Bibliography in BiB-TeX Format
Mladen Victor Wickerhauser
|
|
@Article(fmwwk,
Author = {Nikki L. B. Freeman and Rashmi Muthukkumar and
Ruth S. Weinstock and M. Victor Wickerhauser and
Anna R. Kahkoska},
Title = {Use of machine learning to identify characteristics
associated with severe hypoglycemia in older adults with
type 1 diabetes: a post-hoc analysis of a case-control study},
Journal = {BMJ Open Diabetes Research and Care},
Volume = {12},
Number = {1},
Pages = {1--9},
DOI = {10.1136/bmjdrc-2023-003748},
URL = {http://www.math.wustl.edu/~victor/papers/fmwwk.pdf},
Year = {2024})
@Book(w:finmath,
Author = {Mladen Victor Wickerhauser},
Title = {Introducing Financial Mathematics: Theory, Binomial Models,
and Applications},
URL = {http://www.math.wustl.edu/~victor/finmath},
ISBN = {9781032359854},
Abstract = {Introducing Financial Mathematics: Theory, Binomial
Models, and Applications seeks to replace existing
books with a rigorous stand-alone text that covers
fewer examples in greater detail with more
proofs. The book uses the fundamental theorem of
asset pricing as an introduction to linear algebra
and convex analysis. It also provides example
computer programs, mainly Octave/MATLAB functions
but also spreadsheets and Macsyma scripts, with
which students may experiment on real data.The
text's unique coverage is in its contemporary
combination of discrete and continuous models to
compute implied volatility and fit models to market
data. The goal is to bridge the large gaps among
nonmathematical finance texts, purely theoretical
economics texts, and specific software-focused
engineering texts.},
Publisher = {Chapman and Hall/CRC Press},
Address = {Boca Raton, Florida},
Pages = {xi + 292},
Month = {November},
Year = {2022})
@InCollection(zw:wtnnl,
Author = {Wei Zhu and Mladen Victor Wickerhauser},
Title = {Wavelet Transforms by Nearest Neighbor Lifting},
Abstract = {We show that any discrete wavelet transform using finite
impulse response filters may be factored into lifting
steps that use only nearest-neighbor array elements.
We then discuss the advantages and disadvantages of
imposing this additional requirement.},
URL = {http://www.math.wustl.edu/~victor/papers/zwnnlift.pdf},
DOI = {10.1007/978-0-8176-8379-5_9},
BookTitle = {Excursions in Harmonic Analysis, Volume 2},
Editor = {Travis Andrews and Radu V. Balan and John J. Benedetto
and Wojciech Czaja and Kasso Okoudjou},
Publisher = {Springer-Verlag},
Address = {New York},
Pages = {173--192},
Month = {November},
Year = {2012})
@Article(hmmwmwlw:ispdcumitn,
Author = {Michael Hughes and Jon N. Marsh and John E. McCarthy and
Mladen Victor Wickerhauser and Brian Maurizi and
Kirk D. Wallace and Gregory M. Lanza and Samuel A. Wickline},
Title = {Improved Signal Processing to Detect Cancer by Ultrasonic
Molecular Imaging of Targeted Nanoparticles},
Abstract = {In several investigations of molecular
imaging of angiogenic neovasculature using a targeted
contrast agent, R\'enyi entropy [If(r)] and a limiting
form of R\'enyi entropy (If,infinity) exhibited significantly
more sensitivity to subtle changes in scattering
architecture than energy-based methods. Many of these
studies required the fitting of a cubic spline to
backscattered waveforms prior to calculation of
entropy [either If(r) or If,infinity]. In this study, it is
shown that the robustness of If,infinity may be improved by
using a smoothing spline. Results are presented
showing the impact of different smoothing
parameters. In addition, if smoothing is preceded by
low-pass filtering of the waveforms, further
improvements may be obtained.},
URL = {http://www.math.wustl.edu/~victor/papers/ispdcumitn.pdf},
DOI = {10.1121/1.3578459},
Journal = {Journal of the Acoustical Society of America},
Volume = {129},
Number = {6},
Pages = {3756--3767},
Month = {June},
Year = {2011})
@Article(hammmwww:ussabuwamstdm,
Author = {Michael S. Hughes and Kwesi Agyem and Jon N. Marsh
and John E. McCarthy and Brian N. Maurizi
and M. Victor Wickerhauser and Kirk D. Wallace
and Samuel A. Wickline},
Title = {Use of Smoothing Splines for Analysis of Backscattered
Ultrasonic Waveforms: Application to Monitoring of Steroid
Treatment of Dystrophic Mice},
Abstract = {Duchenne muscular dystrophy (DMD) is an X-linked genetic
disease characterized by progressive weakness and
wasting of skeletal and cardiac muscle; boys present
with weakness by the age of 5 years and, if left
untreated, are unable to walk without assistance by
the age of 10 years. Therapy for DMD has been
primarily palliative, with oral steroids emerging as a
first-line approach even though this treatment has
serious side-effects. Consequently, low-cost imaging
technology suitable for improved diagnosis and
treatment monitoring of DMD would be of great value,
especially in remote and underserved areas.
Previously, we reported use of the logarithm of
the signal energy, log[Ef], and a new method for
ultrasound signal characterization using entropy, Hf,
to monitor prednisolone treatment of skeletal muscle
in a dystrophin-deficient mouse model. Three groups
were studied: mdx mice treated with prednisolone, a
control group of mdx mice treated with saline, and a
control group of wild-type mice treated with
saline. It was found that both log[Ef] and Hf were
required to statistically differentiate the three
groups. In the current study, we show that
preprocessing of the raw ultrasound using optimal
smoothing splines before computation of either
log[Ef] or a rapidly computable variant of Hf,
denoted If,infinity, permits delineation of all three
groups by either metric alone. This opens the way to the
ultimate goal of this study, which is identification
and implementation of new diagnostically sensitive
algorithms on the new generation of low-cost hand-held
clinical ultrasonic imaging systems.},
DOI = {10.1109/TUFFC.2011.2093},
URL = {http://www.math.wustl.edu/~victor/papers/ussabuwamstdm.pdf},
Journal = {IEEE Transactions on Ultrasonics, Ferroelectrics,
and Frequency Control},
Pages = {2361--2369},
Volume = {58},
Number = {11},
Month = {November},
Year = {2011})
@Article(mwmwmlwh:artclfremit,
Author = {Jon N. Marsh and Kirk D. Wallace and John E. McCarthy
and M. Victor Wickerhauser and Brian N. Maurizi and
Gregory M. Lanza and Samuel A. Wickline and Michael S. Hughes},
Title = {Application of a Real-Time, Calculable Limiting Form of the
{R}{\'e}nyi Entropy for Molecular Imaging of Tumors},
Abstract = {Previously, we reported new methods for
ultrasound signal characterization using entropy, Hf;
a generalized entropy, the R\'enyi entropy, If(r); and
a limiting form of R\'enyi entropy suitable for
real-time calculation, If,infinity. All of these quantities
demonstrated significantly more sensitivity to subtle
changes in scattering architecture than energy-based
methods in certain settings. In this study, the
real-time calculable limit of the R\'enyi entropy,
If,infinity, is applied for the imaging of angiogenic murine
neovasculature in a breast cancer xenograft using a
targeted contrast agent. It is shown that this
approach may be used to reliably detect the
accumulation of targeted nanoparticles at five
minutes post-injection in this in vivo model.},
DOI = {10.1109/tuffc.2010.1630},
URL = {http://www.math.wustl.edu/~victor/papers/artclfremit.pdf}
Journal = {IEEE Transactions on Ultrasonics, Ferroelectrics,
and Frequency Control},
Volume = {57},
Number = {8},
Pages = {1890--1895},
Month = {August},
Year = {2010})
@InProceedings(mwlwhmw:alfremitucrp,
Author = {Jon N. Marsh and Kirk D. Wallace and Gregory M. Lanza
and Samuel A. Wickline and Michael S. Hughes and
John E. McCarthy and M. Victor Wickerhauser},
Title = {Application of a Limiting Form of the {R}{\'e}nyi Entropy
for Molecular Imaging of Tumors Using a Clinically Relevant
Protocol},
Abstract = { In earlier studies we reported on the application of
R\'enyi entropy, I_f(r), r<2, for the detection of
precancerous lesions, by detection of subtle changes
in backscattered waveforms f(t) that occured as
targeted nanoparticles slowly accumulated near the
lesion. In contrast, the signal energy defined as the
sum of squares of the signal over the same moving
window, was unable to detect this change (as was
conventional B-mode imaging). Although the
computational effort to obtain the result precluded
its clinical application with currently available
equipment, the study raised the possibility of
further sensitivity improvements by using values of r
closer to the limiting value of 2, where I_f(r)
approaches infinity. The current study demonstrates
that by extracting the asymptotic form of I_f(r) as r
tends to 2 there is no loss of sensitivity. However,
the resulting algorithm is must faster than previous
approaches and has an operation count that is
suitable for implementation in a real-time imaging
system. },
DOI = {10.1109/ULTSYM.2010.5935829},
URL = {http://www.math.wustl.edu/~victor/papers/alfremitucrp.pdf},
BookTitle = {International Ultrasonics Symposium (IUS)},
Institution = {IEEE},
Address = {San Diego, California},
Pages = {53--56},
Month = {11--14 October},
Year = {2010})
@Article(rtrepmiutn:wmwmlwh,
Author = {Kirk Wallace and John McCarthy and Victor Wickerhauser
and Jon Marsh and Gregory Lanza and Samuel Wickline
and Michael Hughes},
Title = {Real-Time {R}enyi Entropy Processing for Molecular Imaging
Using Targeted Nanoparticles},
Abstract = {Previously, improvements in in vivo molecular imaging
sensitivity were obtained using Renyi entropy, Ifr with
values of r near 2, specifically r1.99. This result raised
the possibility of further improvements in sensitivity even
closer to the limit r-->2 at r2, Ifr is undefined. However,
such an investigation was not feasible due to excessive
computational time required to calculate Ifr near this
limit. In this study, an asymptotic expression for the
limiting behavior of Ifr as r-->2 is derived and used to
present results analogous to those obtained with
If1.99. Moreover, the limiting form, If, is computable
directly from the experimentally measured waveform, ft by an
algorithm suitable for real-time implementation. To test our
approach, five mice were injected with 3-targeted
nanoparticles, and ultrasound images obtained at 0-, 15-,
30-, and 45-min post-injection. Two control groups N5,
injected with untargeted-nanoparticles, or no injection were
also imaged. Renyi images were able to differentiate the
groups p0.05 at 15 min post-injection. This outcome agrees
with previous studies using targeted-nanoparticles and
demonstrates the ability of entropy-based signal receivers
when used in conjunction with targetednanoparticles to
elucidate the presence of 3-integrins in primordial
neovasculature, particularly in acoustically unfavorable
environments.}
URL = {http://www.math.wustl.edu/~victor/papers/rtrepmiutn.pdf},
DOI = {10.1121/1.3248766},
Journal = {Journal of the Acoustical Society of America},
Volume = {126},
Number = {4},
Pages = {2214--2214},
Month = {October},
Year = {2009})
@Article(hmwmafwtsalw:rtclfreadscsa,
Author = {Michael S. Hughes and John E. McCarthy
and M. Victor Wickerhauser and Jon N. Marsh
and Jeffery M. Arbeit and Ralph W. Fuhrhop
and Kirk D. Wallace and Lewis Thomas and James Smith
and Kwesi Agyem and Gregory M. Lanza and S. A. Wickline},
Title = {Real-time Calculation of a Limiting Form of the {R}enyi
Entropy Applied to Detection of Subtle Changes in
Scattering Architecture},
Abstract = {Previously we reported a new method for ultrasound
signal characterization using entropy, $H_f$, and
demonstrated that in certain settings, further
improvements in signal characterization could be
obtained by generalizing to Renyi Entropy-based
signal characterizations, $I_f(r)$, with values of
$r$ near 2 (specifically $r=1.99$). We speculated
that further improvements in sensitivity might be
realized at the limit $r\to 2$. At that time, such
investigation was not feasible due to excessive
computational time required to calculate $I_f(r)$
near this limit. In this paper, we now derive an
asymptotic expression for the limiting behavior of
$I_f(r)$ as $r\to 2$ and present results analogous
to those obtained with $I_f(1.99)$. Morover, the
limiting form, $I_{f,\infty}$, is computable directly
from the experimentally measured waveform, $f(t)$, by
an algorithm that is suitable for real-time calculation
and implmentation. },
URL = {http://www.math.wustl.edu/~victor/papers/hmwmafwtsalw.pdf},
DOI = {10.1121/1.3224714},
Journal = {Journal of the Acoustical Society of America},
Volume = {126},
Number = {5},
Pages = {2350--2358},
Month = {November},
Year = {2009})
@InProceedings(ow:kldmommslspp,
Author = {Peter Fogh Odgaard and Mladen Victor Wickerhauser},
Title = {{K}arhunen-{L}o\'{e}ve ({PCA}) based detection of
multiple oscillations in multiple measurement signals
from large-scale process plants},
Abstract = {In the perspective of optimizing the control and operation
of large scale process plants, it is important to detect and
to locate oscillations in the plants. This paper presents a
scheme for detecting and localizing multiple oscillations in
multiple measurements from such a large-scale power
plant. The scheme is based on a Karhunen-Lo\`{e}ve
analysis of the data from the plant. The proposed scheme is
subsequently tested on two sets of data: a set of synthetic
data and a set of data from a coal-fired power plant. In
both cases the scheme detects the beginning of the
oscillation within only a few samples. In addition the
oscillation localization has also shown its potential by
localizing the oscillations in both data sets.},
URL = {http://www.math.wustl.edu/~victor/papers/kloscinew.pdf},
DOI = {10.1109/ACC.2007.4282149},
BookTitle = {Proceedings of the {A}merican {C}ontrol {C}onference 2007},
Publisher = {American Automatic Control Council ({AACC})},
Pages = {5893--5898},
Address = {New York, NY},
Month = {11--13 July},
Year = {2007})
@Article(osawm:fbhsfcdp,
Author = {Peter Fogh Odgaard and Jakob Stoustrup and Palle Andersen
and Mladen Victor Wickerhauser and Henrik Fl\o e Mikkelsen},
Title = {Feature Based Handling of Surface Faults in Compact Disc
Players},
Abstract = { Compact Disc Players have been on the market for
more than two decades and a majority of the
control problems involved have been solved.
However, there are still problems with playing
Compact Discs related to surface faults like
scratches and fingerprints. two servo control
olops are used to keep the Optical Pick-up Unit
focused on the information track of the Compact
Disc. The problem is to design servo controllers
which are well suited for handling surface faults
that disturb position measurements, yet still
react sufficiently against normal disturbances
like mechanical shocks. In this paper a novel
method called feature based control is presented.
The method is based on a fault tolerant control
scheme, which uses extracted features of the
surface faults to remove those from the detector
signals used for control during the occurence of
surface faults. The extracted features are
Karhunen--Lo\`eve approximations of the surface
faults. The performance of the feature based
control scheme is validated by experimental work
with Compact Discs having known surface defects. },
URL = {http://www.math.wustl.edu/~victor/papers/osawm.pdf},
DOI = {10.1016/j.conengprac.2006.01.002},
Journal = {Control Engineering Practice},
Volume = {14},
Number = {12},
Pages = {1495--1509},
Institution = {Washington University},
Address = {Saint Louis, Missouri},
Month = {December},
Year = {2006})
@InProceedings(ow:fpccdp,
Author = {Peter Fogh Odgaard and Mladen Victor Wickerhauser},
Title = {Fault Predictive Control of Compact Disk Players},
Abstract = {Optical disc players such as CD-players have problems
playing certain discs with surface faults like scratches and
fingerprints. The problem is to be found in the servo
controller which positions the optical pick-up, such that
the laser beam is focused on the information track. A scheme
handling this problem, called feature based control, has
been presented in an other publications of the first author.
This scheme is based on an assumption that the surface
faults do not change from encounter to encounter. This
assumption is unfortunately not entirely true. This paper
proposes an improvement of the feature based control scheme,
such that a prediction step is included. The proposed scheme
is compared with the feature based control scheme, in the
perspective of handling surface faults, by
simulations. These simulations show the improvements given
by the proposed algorithm.},
URL = {http://www.math.wustl.edu/~victor/papers/fpccdp.pdf},
DOI = {10.3182/20060829-4-CN-2909.00167},
BookTitle = {Proceedings of 6th {IFAC} Symposium on Fault Detection,
Supervision and Safety of Technical Processes.},
Volume = {39},
Number = {13},
Publisher = {IFAC},
Address = {Beijing, China},
Month = {30 August to 1 September},
Pages = {1063--1068},
Year = {2006})
@InProceedings(osw:wpbdsfcd,
Author = {Peter Fogh Odgaard and Jakob Stoustrup
and Mladen Victor Wickerhauser},
Title = {Wavelet Packet based Detection of Surface Faults
on Compact Discs},
Abstract = {In this paper the detection of faults on the surface of a
compact disc is addressed. Surface faults like scratches
and fingerprints disturb the on-line measurement of the
pick-up position relative to the track. This is critical
since the pick-up is focused on and tracked at the
information track based on these measurements. A precise
detection of the surface fault is a prerequisite to a
correct handling of the faults in order to protect the
pick-up of the compact disc player from audible track
losses. The actual fault handling which is addressed in
other publications can be carried out by the use of
dedicated filters adapted to remove the faults from the
measurements. In this paper detection using wavelet packet
filters is demonstrated. The filters are designed using the
joint best basis method. Detection using these filters shows
a distinct improvement compared to detection using ordinary
threshold methods.},
URL = {http://www.math.wustl.edu/~victor/papers/wpbdsfcd.pdf},
DOI = {10.3182/20060829-4-CN-2909.00184},
BookTitle = {Proceedings of 6th {IFAC} Symposium on Fault Detection,
Supervision and Safety of Technical Processes.},
Volume = {39},
Number = {13},
Publisher = {IFAC},
Address = {Beijing, China},
Month = {30 August to 1 September},
Pages = {1103--1108},
Year = {2006})
@InCollection(w:introix,
Author = {Mladen Victor Wickerhauser},
Title = {Introduction to {S}ection {IX}: Selected Applications},
URL = {http://www.math.wustl.edu/~victor/papers/introix.pdf},
Abstract = {Over the past decade, wavelet transforms have been widely
applied. Good implementations of the discrete wavelet
transform (DWT) were built into software systems such as
Matlab and S-Plus, and DWT became a frequently-used tool for
data analysis and signal processing. There are certain
problems, though, on which this tool works particularly
well. The most common ingredient in those problems is some
complicated object that can be closely approximated by a few
superposed wavelets. This compilation includes four seminal
articles that introduced some of these stand-out DWT
applications. I have taken a random and sparse sampling of
relevant articles and books published around the same time,
in order to place the results in context and illustrate
their influence.},
Pages = {733--740},
BookTitle = {Fundamental Papers in Wavelet Theory},
ISBN = {cloth:0-691-11453-6,paper:0-691-12705-0},
Editor = {Christopher Heil and David Walnut},
Publisher = {Princeton University Press},
Address = {Princeton, New Jersey},
Month = {July},
Year = {2006})
@Article(elnetal:cmcdcdt,
Title = {A Comparison of {M}onte {C}arlo Dose Calculation
Denoising Techniques},
Author = {I. El Naqa and I. Kawrakow and M. Fippel and J. V. Siebers
and P. E. Lindsay and M. V. Wickerhauser and M. Vicic and
K. Zakarian and N. Kauffmann and J. O. Deasy},
Abstract = {Recent studies have demonstrated that Monte Carlo (MC)
denoising techniques can reduce MC radiotherapy dose
computation time significantly by preferentially
eliminating statistical fluctuations (`noise') through
smoothing. In this study, we compare new and previously
published approaches to MC denoising, including 3D
wavelet threshold denoising with sub-band adaptive
thresholding, content adaptive mean-median-hybrid
(CAMH) filtering, locally adaptive Savitzky--Golay
curve-fitting (LASG), anisotropic diffusion (AD) and an
iterative reduction of noise (IRON) method formulated
as an optimization problem. Several challenging phantom
and computed-tomography-based MC dose distributions
with varying levels of noise formed the test
set. Denoising effectiveness was measured in three
ways: by improvements in the mean-square-error (MSE)
with respect to a reference (low noise) dose
distribution; by the maximum difference from the
reference distribution and by the `Van Dyk' pass/fail
criteria of either adequate agreement with the
reference image in low-gradient regions (within 2
percent in our case) or, in high-gradient regions, a
distance-to-agreement-within-2-percent of less than 2
mm. Results varied significantly based on the dose
test case: greater reductions in MSE were observed for
the relatively smoother phantom-based dose distribution
(up to a factor of 16 for the LASG algorithm); smaller
reductions were seen for an intensity modulated
radiation therapy (IMRT) head and neck case (typically,
factors of 2.4). Although several algorithms reduced
statistical noise for all test geometries, the LASG
method had the best MSE reduction for three of the four
test geometries, and performed the best for the Van Dyk
criteria. However, the wavelet thresholding method
performed better for the head and neck IMRT geometry
and also decreased the maximum error more effectively
than LASG. In almost all cases, the evaluated methods
provided acceleration of MC results towards
statistically more accurate results.},
URL = {http://www.math.wustl.edu/~victor/papers/cmcdcdt.pdf},
DOI = {10.1088/0031-9155/50/5/014},
Journal = {Physics in Medicine and Biology},
Volume = {50},
Year = {2005},
Pages = {909--922})
@Article(cmwj:fwewb,
Author = {Elvir \v{C}au\v{s}evi\'c and Robert E. Morley and
M. Victor Wickerhauser and Arnaud E. Jacquin},
Title = {Fast Wavelet Estimation of Weak Biosignals},
Abstract = {Wavelet based signal processing has become commonplace in
the signal processing community over the past decade and
wavelet based software tools and integrated circuits are now
commercially available. One of the most important
applications of wavelets is in removal of noise from
signals, called denoising, accomplished by thresholding
wavelet coefficients in order to separate signal from
noise. Substantial work in this area was summarized by
Donoho and colleagues at Stanford University, who developed
a variety of algorithms for conventional denoising.
However, conventional denoising fails for signals with low
signal-to-noise ratio (SNR). Electrical signals acquired
from the human body, called biosignals, commonly have below
0 dB SNR. Synchronous linear averaging of a large number of
acquired data frames is universally used to increase the SNR
of weak biosignals. A novel wavelet-based estimator is
presented for fast estimation of such signals. The new
estimation algorithm provides a faster rate of convergence
to the underlying signal than linear averaging. The
algorithm is implemented for processing of auditory
brainstem response (ABR) and of auditory middle latency
evoked potential response (AMLR) signals. Experimental
results with both simulated data and human subjects
demonstrate that the novel wavelet estimator achieves
superior performance to that of linear averaging.},
URL = {http://www.math.wustl.edu/~victor/papers/cmwj.pdf},
DOI = {10.1109/TBME.2005.846722},
Journal = {IEEE Transactions on Biomedical Engineering},
Month = {June},
Volume = {52},
Number = {6},
Pages = {1021--1032},
Month = {May},
Year = {2005})
@TechReport(gw:ptcnii,
Author = {William F. Gossling and Mladen Victor Wickerhauser},
Title = {Prices, the Trade Cycle, and the Nature of
Industrial Interdependence},
Abstract = {The continuance of a rise in prices in a Western economy
well after a downturn in final demands, termed
``Stagflation,'' has been a puzzle to economists in the
20th century: we hope, from the results set out in this
paper, that any stagflation encountered in the 21st
century will have been understood and even anticipated (in
the proper sense of that word) by appropriate economic
policies which include the use of input-output (or
interindustry) tables. In conclusion, at the end of this
paper, attention is drawn to the computability of
projections of industrial price-levels and rates of return:
the ``duals'' to the familiar ``primal'' projections
(embracing industrial outputs and growth rates) into the
future. This leads one to a conjoint ``forward view'' (see
Gielnik, 1980, in ``Input, Output, and Marketing,''
London, I.-O. P. C.) which should be applicable at least to
most economies which have enough of the required data.},
URL = {http://www.math.wustl.edu/~victor/papers/gw.pdf},
Software = {http://www.math.wustl.edu/~victor/software/cycles/index.html},
Pages = {13},
Institution = {Washington University},
Address = {Saint Louis, Missouri},
Year = {2004})
@TechReport(osawm:smfrs,
Author = {Peter Fogh Odgaard and Jakob Stoustrup and Palle Andersen
and Mladen Victor Wickerhauser and Henrik Fl\o e Mikkelsen},
Title = {A Simulation Model of Focus and Radial Servos in
{C}ompact {D}isc players with Disc Surface Defects},
URL = {http://www.math.wustl.edu/~victor/papers/cdsim.pdf},
Abstract = {Compact Disc players have been on the market for more
than two decades, and most of the control servo
problems have been solved. One large remainig
problem is the handling of severe surface defects
like scratches and fingerprints. This paper
introduces a method for making the design of
controllers handling surface defects easier: a model
simulating a Compact Disc player reading a disc with
surface defects. The model is based on data from
discs with known surface defects. It is used to
compare a high-bandwidth and a low-bandwidth
controller's performance handling surface defects.},
Note = {Accepted by the {\em Proceedings of the IEEE Joint CCA,
ISIC and CACSD}, September 2--4, 2004, Taipei, Taiwan.},
Pages = {8},
Institution = {Washington University},
Address = {Saint Louis, Missouri},
Year = {2003})
@Misc(w:mfmm-ex,
Title = {Additional Solved Exercises from
{\em {M}athematics for {M}ultimedia}},
Author = {Mladen Victor Wickerhauser},
HowPublished = {Available to instructors from Elsevier's Faculty Lounge},
Note = {236 pages},
Month = {September},
Year = {2004})
@InProceedings(wc:spie5439,
Author = {Mladen Victor Wickerhauser and Wojciech Czaja},
Title = {A Simple Nonlinear Filter for Edge Detection in Images},
Abstract = {We specialize to two simple cases the algorithm for
singularity detection in images from eigenvalues of the dual
local autocovariance matrix. The eigenvalue difference, or
``edginess'' at a point, then reduces to a simple nonlinear
function. We discuss the derivation of these functions,
which provide low-complexity nonlinear edge filters with
parameters for customization, and obtain formulas in the two
simplest special cases. We also provide an implementation
and exhibit its output on six sample images.},
URL = {http://www.math.wustl.edu/~victor/papers/mpedge.pdf},
Software = {http://www.math.wustl.edu/~victor/papers/mpedge.zip},
CrossRef = {spie5439},
Pages = {24--31},
Year = {2004})
@Proceedings(spie5439,
Editor = {Harold H. Szu and Mladen V. Wickerhauser
and Barak A. Pearlmutter and Wim Sweldens},
Title = {Independent Component Analyses, Wavelets, Smart Sensors,
and Neural Networks {II}},
BookTitle = {Independent Component Analyses, Wavelets, Smart Sensors,
and Neural Networks {II}},
Publisher = {SPIE},
Series = {SPIE Proceedings},
Volume = {5439},
Address = {Orlando, Florida},
ISBN = {0-8194-5362-5},
ISSN = {0277-786X},
Month = {14--15 April},
Year = {2004})
@Book(w:mfmm,
Author = {Mladen Victor Wickerhauser},
Title = {Mathematics for Multimedia},
URL = {http://www.math.wustl.edu/~victor/mfmm},
ISBN = {0-12-748451-5},
Abstract = {This textbook presents selected results in algebra and
analysis, chosen for their usefulness in understanding and
creating application software for multimedia signal
processing and communication. Over one hundred exercises
with complete solutions as well as example programs in
Standard C are included to aid the student. The material
forms the basis for a one-semester upper-level undergraduate
course of Topics in Applied Mathematics at Washington
University.},
Publisher = {Elsevier/Academic Press},
Address = {San Diego, California},
Pages = {xii + 302},
Month = {November},
Year = {2003})
@TechReport(ow:discr,
Author = {Peter Fogh Odgaard and Mladen Victor Wickerhauser},
Title = {Discrimination Between Different Kinds of Surface Defects
on Compact Discs},
URL = {http://www.math.wustl.edu/~victor/papers/odgaardd.pdf},
Abstract = {Compact Disc players have problems playing discs with surface
defects such as scratches and finger prints. The problem
is that handling normal disturbances such as mechanical
shocks requires a high bandwidth of the controllers that
keep the Optical Pick-Up focused and radially placed on the
information track on the disc. In order for the
controllers to handle the surface defects it is required
that they are non-sensitive to the frequency contents of
the defect, since a defect can be viewed as a disturbance
on the measurements. A simple solution to this problem is
to decrease the controller bandwidth during the defect.
However, due to the variation of defects a more adaptive
control strategy would be preferable. In this paper the
defects are categorised into three groups. A discriminator
is designed, based on the local most-discriminating basis
vectors of the Karhunen-Lo\`eve and Haar bases as well as
the mean of defect group vectors. In these bases the
discrimination rule is simple. The defect in question is a
member of the group it is closest to. The Karhunen-Lo\`eve
basis gives a correct classification rate of more than 85.7
percent with three basis vectors and the Haar basis of
more than 94.6 percent with 5 basis vectors.},
Note = {Accepted by the {\em Proceedings of IECON 2004}, Busan, Korea},
Pages = {6},
Institution = {Washington University},
Address = {Saint Louis, Missouri},
Year = {2003})
@TechReport(ow:timeloc,
Author = {Peter Fogh Odgaard and Mladen Victor Wickerhauser},
Title = {Time Localisation of Surface Defects on Optical Discs},
URL = {http://www.math.wustl.edu/~victor/papers/odgaardt.pdf},
Abstract = {Many have experienced problems with their Compact Disc
player when a disc with a scratch or finger print is
played. One way to improve the playability of discs with
such defects is to locate the defect in time and then
handle it in a special way. This time localisation must
be quite accurate, and Fang's algorithm for segmentation of
the time axis is used because of its good performance in
similar applications. For those defects where Fang's
algorithm fails, the usual variance threshold method may be
used as it handles eccentricity and end localisation
better. },
Note = {Accepted by the {\em Proceedings of the IEEE Joint CCA,
ISIC and CACSD}, September 2--4, 2004, Taipei, Taiwan.},
Pages = {6},
Institution = {Washington University},
Address = {Saint Louis, Missouri},
Year = {2003})
@Proceedings(spie5102,
Editor = {Anthony J. Bell and Mladen V. Wickerhauser
and Harold H. Szu},
Title = {Independent Component Analyses, Wavelets
and Neural Networks},
BookTitle = {Independent Component Analyses, Wavelets
and Neural Networks},
Publisher = {SPIE},
Series = {SPIE Proceedings},
Volume = {5102},
Address = {Orlando, Florida},
ISBN = {0-8194-4962-8},
ISSN = {0277-786X},
Month = {22--25 April},
Year = {2003})
@Article(w:ajse,
Author = {Mladen Victor Wickerhauser},
Title = {Some Problems Related to Wavelet Packet Bases
and Convergence},
URL = {http://www.math.wustl.edu/~victor/papers/ajse.pdf},
Abstract = {Wavelet packets are subsets of a multiresolution analysis
and derive many of their properties therefrom. Those
defined by a single filter pair have uncontrolled size and
basis properties, in general. By substituting different
filters at different scales according to a rule, these can
be controlled. The number of orthonormal bases available in
an MRA satisfies a recursion equation depending on the basis
selection method, and some of these recursions have closed
form solutions. Some of these orthonormal bases consist of
uniformly bounded, uniformly compactly supported wavelet
packets and are Schauder bases for many Banach spaces. With
controlled size and support, the Carleson--Hunt theorem
applies to show that a wavelet packet Fourier series of a
continuous function converges pointwise almost everywhere.},
Journal = {Arabian Journal for Science and Engineering},
Volume = {28},
Number = {1C},
Pages = {45-58},
Month = {June},
Year = {2003})
@Article(w:ripples,
URL = {http://www.math.wustl.edu/~victor/papers/ripples.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Two Introductions to Wavelets},
Abstract = {Review of {\em Ripples in Mathematics: The Discrete
Wavelet Transform}, by A. Jensen and A. la
Cour-Harbo, ISBN 3-540-41662, and {\em A First Course
in Wavelets with Fourier Analysis}, by Albert Boggess
and Francis J. Narcowich, ISBN 0-13-022809-5},
Note = {Book review.},
Pages = {163--167},
Journal = {American Mathematical Monthly},
Volume = {110},
Number = {2},
Month = {February},
Year = {2003})
@Misc(w:swaa,
URL = {http://www.math.wustl.edu/~victor/talks/mvwswaa1.pdf,
http://www.math.wustl.edu/~victor/talks/mvwswaa2.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Survey of Wavelet Algorithms and Applications},
HowPublished = {SPIE Short Course Notes SC475, AeroSense 2002,
Orlando, Florida},
Abstract = {DESCRIPTION: A brief description of wavelets and wavelet
packets will be followed by a moderately detailed survey of
fast discrete wavelet transform algorithms and
implementations. Emphasis will be placed on the ``lifting''
implementation, treatment of boundaries, and wavelet and
basis selection, keyed to the transforms used in the WSQ and
JPEG2000 image compression algorithms. The related lapped
orthogonal transforms will be discussed as well.
LEARNING OUTCOMES: At the conclusion, students should have
learned to:
- define wavelets and wavelet packets and state
their useful mathematical properties.
- implement a basic wavelet transform by the lifting
method;
- implement one or more boundary treatments for
wavelet transforms;
- choose an appropriate wavelet for a given signal
class;
- describe the transforms used in the WSQ and
JPEG2000 image compression algorithms.
INTENDED AUDIENCE: The lecture will be pitched to engineers
with some programming and signal processing experience, but
will be understandable to anyone with a basic undergraduate
mathematics, science, or engineering preparation. It will
focus on the image compression example, but is intended to
be useful to anyone seeking deeper understanding of the
mathematical principles underlying signal analysis.
},
Month = {April 4},
Year = {2002})
@Article(w:jmr,
URL = {http://www.math.wustl.edu/~victor/papers/jmr.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {{\bf Wavelets: Tools for Science \& Technology},
by St\'ephane Jaffard, Yves Meyer, and Robert D. Ryan},
Abstract = {Review of {\em Wavelets: Tools for Science \& Technology},
by St\'ephane Jaffard, Yves Meyer, and Robert D. Ryan, SIAM,
Philadelphia, 2001, ISBN 0-89871-448-6.},
Pages = {302--305},
Journal = {SIAM Review},
Volume = {44},
Number = {2},
Year = {2002})
@InProceedings(w:pwaa,
URL = {http://www.math.wustl.edu/~victor/papers/pwaa.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Progress in Wavelet Algorithms and Applications},
Abstract = {Wavelet and wavelet packet transforms are presently used for
image compression and denoising. There has been recent
progress on three fronts: implementing multiplication
operations in wavelet bases, estimating compressibility by
wavelet packet transform coding, and designing wavelet
packets to control frequency spreading and pointwise
convergence. Some open problems are mentioned.},
Editor = {Harold H. Szu},
BookTitle = {Wavelet and Independent Component Analysis
Applications {IX}},
Publisher = {SPIE},
Series = {SPIE Proceedings},
Volume = {4738},
Address = {Orlando, Florida},
Month = {3--5 April},
CrossRef = {spie4738},
Pages = {157--168},
Year = {2002})
@InProceedings(w:aec,
URL = {http://www.math.wustl.edu/~victor/papers/aec.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Basis and Convergence Properties of Wavelet Packets},
Abstract = {Wavelet packets defined by a single filter pair have
uncontrolled size and basis properties, in general. By
substituting different filters at different scales
according to a rule, these can be controlled. One can
obtain Schauder bases of uniformly bounded, uniformly
compactly supported wavelet packets. By controlling size
and support, one can apply the Carleson--Hunt theorem to
show that certain wavelet packet Fourier series of a
continuous function converge almost everywhere.},
BookTitle = {Proceedings of the International Conference on Wavelet
Analysis and Applications, Guangzhou, China, November, 1999},
Editor = {Donggao Deng and Daren Huang and Rong-Qing Jia and Wei Lin
and Jianzhong Wang},
Publisher = {American Mathematical Society, International Press},
Address = {Providence, Rhode Island},
Series = {AMS/IP Studies in Advanced Mathematics},
Volume = {25},
Pages = {279-287},
ISBN = {0-8218-2991-2},
Year = {2002})
@InProceedings(w:awaa,
URL = {http://www.math.wustl.edu/~victor/papers/awaa.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Advances in Wavelet Algorithms and Applications},
Abstract = {Wavelet and wavelet packet transforms are presently used for
image compression and denoising. There has been recent
progress on three fronts: implementing multiplication
operations in wavelet bases, estimating compressibility by
wavelet packet transform coding, and designing wavelet
packets to control frequency spreading and pointwise
convergence. Some open problems are mentioned.},
BookTitle = {Wavelet Analysis: Twenty Years' Developments},
Note = {Proceedings of the International Conference on Computational
Harmonic Analysis, City University of Hong Kong, 4--8 June, 2001},
Editor = {Ding-Xuan Zhou},
Publisher = {World Scientific Publishing},
Address = {Singapore},
Pages = {289-310},
ISBN = {981-238-142-2},
Year = {2002})
@Book(twyl:waa01,
Editor = {Yuan Y. Tang and Victor Wickerhauser and Pong C. Yuen
and Chun-hung Li},
Title = {Wavelet Analysis and Its Applications},
Series = {Lecture Notes in Computer Science},
Volume = {2251},
Note = {Proceedings of the Second International Conference, {WAA}
2001, Hong Kong, China, December, 2001},
Publisher = {Springer-Verlag},
Address = {Berlin},
ISBN = {3-540-43034-2},
Year = {2001})
@Article(cw:sdiudla,
Author = {Wojciech Kladiusz Czaja and Mladen Victor Wickerhauser},
Title = {Singularity Detection in Images Using Dual Local
Autocovariance},
URL = {http://www.math.wustl.edu/~victor/papers/edges.pdf},
Software = {http://www.math.wustl.edu/~victor/software/acha/edges-c.zip},
Abstract = { We use the eigenvalues of a version of the autocovariance
matrix to recognize directions at which the Fourier
transform of a function is slowly decreasing, which
provides us with a technique to detect singularities in
images. },
Journal = {Applied and Computational Harmonic Analysis},
Volume = {13},
Number = {1},
Pages = {77--88},
Month = {July},
Year = {2002})
@Article(dwp:amcs,
URL = {http://www.math.wustl.edu/~victor/papers/dwp.pdf},
Author = {Joseph O. Deasy and M. Victor Wickerhauser
and Mathieu Picard},
Title = {Accelerating {M}onte {C}arlo Simulations of
Radiation Therapy Dose Distributions Using
Wavelet Threshold De-Noising},
Abstract = {Limit distributions approximated by long Monte Carlo
simulations can also be obtained with good precision by
applying wavelet threshold denoising to much shorter
simulations.},
Journal = {Medical Physics},
Volume = {29},
Number = {10},
Pages = {2366--2373},
Year = {2002})
@Article(sw:icf,
URL = {http://www.math.wustl.edu/~victor/papers/icf.pdf},
Author = {Hrvoje \v{S}iki\'{c} and Mladen Victor Wickerhauser},
Title = {Information Cost Functions},
Abstract = { We derive some curious inequalities for discrete
probability densities by carefully examining certain
Schur concave functions.},
Journal = {Applied and Computational Harmonic Analysis},
Volume = {11},
Number = {2},
Pages = {147--166},
Month = {September},
Year = {2001})
@TechReport(ww:icisp01,
Author = {Eva Wesfreid and Mladen Victor Wickerhauser},
Title = {Frequency Change Function and Acoustic Signals},
Type = {Preprint},
URL = {http://www.math.wustl.edu/~victor/papers/icisp01.pdf},
Abstract = { The local cosine4 orthonormal bases are particularly well
adapted for analyzing signals with piecewise time
behaviour. There are many acoustic signals in music and
speech processing that can be considered as a sequence of
overlapping elementary structures such as phonemes in
speech signals. The Best Basis algorithm computes a local
spectrum defined over a dyadic segmentation, however, there
is no reason for elementary structures to begin and end
near dyadic points. We use Fang's algorithm which segments
the time axis into intervals of arbitrary length; this
algorithm constructs a frequency change function whose
local maxima denote structure changes. The smooth cosine4
orthonormal basis defined over this segmentation is used to
compute a local spectrum associated with elementary
structures. We show that this representation compared with
the Best Basis coefficients has less reconstruction
distortion and better local pattern description.},
Note = {Proceedings of the First International Conference on Image
and Signal Processing (ICISP 2001)},
Institution = {Ibn Zohr University},
Address = {Agadir, Morocco},
Pages = {7},
Month = {May},
Year = {2001})
@TechReport(w:aec-0,
URL = {http://www.math.wustl.edu/~victor/papers/aec.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Basis and Convergence Properties of Wavelet Packets},
Abstract = {Wavelet packets defined by a single filter pair have
uncontrolled size and basis properties, in general. By
substituting different filters at different scales
according to a rule, these can be controlled. One can
obtain Schauder bases of uniformly bounded, uniformly
compactly supported wavelet packets. By controlling size
and support, one can apply the Carleson--Hunt theorem to
show that certain wavelet packet Fourier series of a
continuous function converge almost everywhere.},
Institution = {Washington University},
Address = {Saint Louis, Missouri},
Type = {Preprint},
Pages = {9},
Month = {10 May},
Year = {2000})
@InCollection(pw:mswsucc,
URL = {http://www.math.wustl.edu/~victor/papers/connec.pdf},
Author = {Valerie Perrier and Mladen Victor Wickerhauser},
Title = {Multiplication of Short Wavelet Series Using
Connection Coefficients},
Abstract = {Given two functions approximable with short wavelet
series, we wish to find the short wavelet series
representing their product. This can be done by
pre-calculating the {\em connection coefficients} which
express the product of two wave\-lets or scaling
functions as a wavelet series. We follow a method
suggested by Daubechies and also used by Dahmen, {\it
et al.}, to rapidly compute these coefficients as
elements of a matrix which solves a fixed-point
problem, and derive some of the formulas and identities
satisfied by the coefficients. We estimate the
complexity of the connection coefficient multiplication
algorithm by counting the number of terms, and then
illustrate through a series of graphs how few of these
terms are non-negligible.},
BookTitle = {Advances in Wavelets},
Editor = {Ka-Sing Lau},
ISBN = {981-4021-08-3},
Publisher = {Springer-Verlag},
Address = {Singapore},
Pages = {77--101},
Year = {1999})
@InProceedings(wwc:cimaf99,
URL = {http://www.math.wustl.edu/~victor/papers/cimaf99.pdf},
Author = {Eva Wesfreid and Mladen Victor Wickerhauser},
Title = {Vocal Command Signal Segmentation and Phoneme
Classification},
Abstract = { We show that Xiang Fang's segmentation algorithm of
nearly constant instantaneous frequency is well-adapted
to some noisy vocal command signals and that the
orthonormal local trigonometric transform defined over
this segmentation offers an optimal, non-dyadic
time-frequency tiling. We use this method to compute a
local spectrum for speech processing corresponding
nearly to phonemes and, in a biomedical application, to
measure velopharyngeal closure timing for a swallowing
sound.},
BookTitle = {Proceedings of the II Artificial Intelligence Symposium at
{CIMAF} 99},
Editor = {Alberto A. Ochoa.},
Organization = {Institute of Cybernetics, Mathematics and Physics (ICIMAF),
Habana, Cuba},
Pages = {10},
Year = {1999})
@InCollection(w:wta,
URL = {http://www.math.wustl.edu/~victor/papers/watc.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Wavelet Transforms},
Abstract = { This article will present a very brief sketch of the
history and theory of wavelet analysis, and then list a
few applications to physical and computational
chemistry.},
Volume = {5},
Pages = {3214--3222},
BookTitle = {Encyclopedia of Computational Chemistry},
Editor = {P. v. R. Schleyer and N. L. Allinger and T. Clark and
J. Gasteiger and P. A. Kollman and Henry F. Schaeffer III
and P. R. Schreiner},
Publisher = {John Wiley \& Sons, Limited},
Address = {Chichester, England},
ISBN = {0-471-96588-X},
Year = {1998})
@InProceedings(w:dcwpicsafsd,
URL = {http://www.math.wustl.edu/~victor/papers/grossman.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Designing a Custom Wavelet Packet Image Compression Scheme,
with Applications to Fingerprints and Seismic Data},
Abstract = { No single image compression algorithm can be expected
to work well for all images, and designing a transform
coding image compression algorithm for a given
application is itself a meta-algorithm. Sampling
rates, frequency content, and pixel quantization all
influence the compressibility of the original data.
Subsequent machine or human analyses of the compressed
data, or its presentation at various magnifications,
all influence the nature and visibility of distortion
and artifacts. Thus, algorithms like JPEG, established
for a ``natural'' images intended to be viewed by
humans, do not satisfy the requirements for compressing
fingerprint images intended to be scanned by machines.
In that particular example, it was necessary to develop
a new algorithm {\em WSQ}.
One procedure focuses on the transform portion of the
compression algorithm: the {\em best basis method\/}
automatically finds a transform which provides the best
average compression of a representative set of images,
selected from a set of ``fast'' transforms. A version
of this method was used to design the WSQ fingerprint
image compression algorithm, while another was used to
design compression algorithms for various types of
seismic exploration data.},
BookTitle = {Perspectives in Mathematical Physics: Conference in honor
of Alex Grossmann},
Editor = {Matthias Holschneider and Ginette Saracco},
Publisher = {CRC Press},
Organization = {CFML},
Address = {Marseille-Luminy, France},
Pages = {5},
Month = {July},
Year = {1998})
@InProceedings(wwb,
URL = {http://www.math.wustl.edu/~victor/papers/wwb.pdf},
Author = {Eva Wesfreid and Mladen Victor Wickerhauser and
R. Bouguerra},
Title = {Well Adapted Non Dyadic Local Spectrum for Some Acoustic
Signals},
Abstract = { We show that Fang's segmentation algorithm of nearly
constant instantaneous frequency is well-adapted to
some noisy vocal command signals and that the
orthonormal trigonometric basis of l^2(Z) defined over
this segmentation offers an optimal, non-dyadic
time-frequency tiling. We use this basis in speech
processing to compute a local spectrum and approximate
phonemes. We also use the algorithm to measure
velopharyngeal closure timings from swallowing sounds
for biomedical applications.},
Pages = {223--225},
BookTitle = {Proceedings of IWC-Tangier 98, International Wavelets
Conference ``Wavelets and Multiscale Methods''},
Editor = {Aline Bonami and Albert Cohen and Abdelhak Ezzine and
Paolo Gon\c calv\`es and St\'ephane Jaffard
and Yves Meyer},
Month = {13--17 April},
Organization = {INRIA, Rocquencourt, France},
Address = {Tangier, Morocco},
Year = {1998})
@InCollection(cw:eawdmsi,
URL = {http://www.math.wustl.edu/~victor/papers/eawdmsi.pdf},
Author = {Ronald Raphael Coifman and Mladen Victor Wickerhauser},
Title = {Experiments with Adapted Wavelet De-Noising for Medical
Signals and Images},
Abstract = { We first describe some new libraries of waveforms,
including wavelets, wavelet packets, and local sines
and cosines, which are well-adapted to representing
biological and biomedical signals. By expanding a
signal in a library of waveforms which are
well-localized in both time and frequency, we can
separate coherent structures from incoherent noise. We
experiment with one implementation of adapted wavelet
denoising, and compute the signal-to-noise ratio
improvement obtained for certain simple signals.},
Pages = {323--346},
BookTitle = {Time-Frequency and Wavelets in Biomedical Engineering},
Editor = {Metin Akay},
Publisher = {IEEE Press},
ISBN = {0-7803-1147-7},
Address = {Piscataway, New Jersey},
Year = {1998})
@InProceedings(klgcw:wbarmdis,
URL = {http://www.math.wustl.edu/~victor/papers/klgcw.pdf},
Author = {Mingqi Kong and Jean-Pierre Leduc and Bijoy K. Ghosh and Jonathan R. Corbett
and Mladen Victor Wickerhauser},
Title = {Wavelet Based Analysis of Rotational Motion in Digital
Image Sequences},
Abstract = { This paper addresses the problem of estimating, analyzing
and tracking objects moving with spatio-temporal rotational
motion (spin or orbit). It is assumed that the digital
signals of interest are acquired from a camera and
structured as digital image sequences. The trajectories in
the signal are two-dimensional spatial projections in time
of motion taking place in a three-dimensional space. The
purpose of this work is to focus on the rotational motion,
i.e. estimate the angular velocity. In natural scenes,
rotational motion usually composes with translational or
accelerated motion on a trajectory. This paper shows that
trajectory parameters and rotational motion can be
efficiently estimated and tracked either simultaneously or
separately. The final goal of this work is to provide
selective reconstructions of moving objects of
interest. This paper constructs new continuous wavelet
transforms that can be tuned to both translational and
rotational motion. The parameters of analysis that are taken
into account in these rotational wavelet transforms are
space and time position, velocity, spatial scale, angular
orientation and angular velocity. The continuous wavelet
functions are finally discretized for signal processing. The
link between rotational motion, symmetry and critical
sampling is also presented. Applications are presented with
tracking and estimation.},
Volume = {5},
Pages = {2777--2780},
BookTitle = {Proceedings of ICASSP-98, Seattle},
Publisher = {IEEE Press},
Organization = {IEEE},
Address = {Piscataway, New Jersey},
Month = {12--15 May},
Year = {1998})
@InProceedings(lckwg:astwt,
URL = {http://www.math.wustl.edu/~victor/papers/lckwg.pdf},
Author = {Jean-Pierre Leduc and Jonathan R. Corbett and Mingqi Kong
and Mladen Victor Wickerhauser and Bijoy K. Ghosh},
Title = {Accelerated Spatio-temporal Wavelet Transforms: An
Iterative Trajectory Estimation},
Abstract = { We estimate and analyze accelerated motion in digital
image sequences, using expansions in a new continuous
wavelet transform. Our wavelets are parametrized by
spatial and temporal position, velocity, and
acceleration, spatial scale and spatial rotation. We
can produce selective reconstructions of accelerated
objects of interest.},
Volume = {5},
Pages = {2781--2784},
BookTitle = {Proceedings of ICASSP-98, Seattle},
Publisher = {IEEE Press},
Organization = {IEEE},
Address = {Piscataway, New Jersey},
Month = {12--15 May},
Year = {1998})
@InProceedings(lcw:rwtmaet,
URL = {http://www.math.wustl.edu/~victor/papers/lcw.pdf},
Author = {Jean-Pierre Leduc and Jonathan R. Corbett
and Mladen Victor Wickerhauser},
Title = {Rotational Wavelet Transforms for Motion Analysis,
Estimation, and Tracking},
Abstract = { We estimate and analyze rotational motion in digital
image sequences, using expansions in a new continuous
wavelet transform. Our wavelets are parametrized by
spatial and temporal position, velocity, and
spatial rotation. The continuous wavelet functions
are ultimately discretized for signal processing.},
BookTitle = {Proceedings of the 1998 IEEE International Conference
on Image Processing (ICIP-98), Chicago, Illinois,
October 4-7, 1998},
Volume = {2},
Pages = {195-199},
ISBN = {0-8186-8821-1},
Publisher = {IEEE Press},
Organization = {IEEE Computer Society},
Address = {Piscataway, New Jersey},
Year = {1998)
@InProceedings(klgw:stcwtmbsris,
URL = {http://www.math.wustl.edu/~victor/papers/klgw.pdf},
Author = {Mingqi Kong and Jean-Pierre Leduc and Bijoy K. Ghosh
and Mladen Victor Wickerhauser},
Title = {Spatio-Temporal Continuous Wavelet Transforms for
Motion-Based Segmentation in Real Image Sequences},
Abstract = { We describe an algorithm to track objects moving with
spatio-temporal rotation in digital image sequences,
using expansions in a new continuous wavelet transform.
We show that trajectory parameters and rotational
motion can be efficiently estimated and tracked. The
link between rotational motion, symmetry, and critical
sampling is also discussed.},
BookTitle = {Proceedings of the 1998 IEEE International Conference
on Image Processing (ICIP-98), Chicago, Illinois,
October 4-7, 1998},
Volume = {2},
Pages = {662-666},
ISBN = {0-8186-8821-1},
Publisher = {IEEE Press},
Organization = {IEEE Computer Society},
Address = {Piscataway, New Jersey},
Year = {1998})
@InProceedings(vw:cwicssdc,
URL = {http://www.math.wustl.edu/~victor/papers/spie3169.pdf},
Author = {Anthony Vassiliou and Mladen Victor Wickerhauser},
Title = {Comparison of Wavelet Image Coding Schemes for Seismic
Data Compression},
Abstract = { Wavelet transform coding image compression is applied
to two raw seismic data sets. The parameters of filter
length, depth of decomposition, and quantization method
are varied through 36 parameter settings and the
rate-distortion relation is plotted and fitted with a
line. The lines are compared to judge which parameter
setting produces the highest quality for a given
compression ratio on the sample data. It is found that
long filters, moderate decomposition depths, and
frequency-weighted, variance-adjusted quantization
yield the best results.},
Editor = {Akram Aldroubi and Andrew F. Laine and Michael A. Unser},
BookTitle = {Wavelet Applications in Signal and Image Processing {V}},
Volume = {3169},
Month = {July},
Publisher = {SPIE},
Organization = {SPIE},
Pages = {9},
Year = {1997})
@InCollection(wfg:tdcst,
URL = {http://www.math.wustl.edu/~victor/papers/tdcst.pdf},
Author = {Mladen Victor Wickerhauser and Marie Farge and Eric Goirand},
Title = {Theoretical Dimension and the Complexity of Simulated
Turbulence},
Abstract = {A global quantity called ``theoretical dimension'' is
roughly proportional to the number of coherent
structures that expert observers count in simulated
two-dimensional turbulent viscous flows. The quantity
is computed for a few academic examples and then for a
small number of flows computed from random initial
vorticity fields.},
Pages = {473--492},
Editor = {Wolfgang Dahmen and Peter Oswald and Andrew J. Kurdila},
BookTitle = {Multiscale Wavlet Methods for Partial Differential
Equations},
Series = {Wavelet Analysis and Applications},
Volume = {6},
Publisher = {Academic Press},
ISBN = {0-12-200675-5},
Address = {Boston},
Year = {1997})
@Article(h-nw:wtfa,
URL = {http://www.math.wustl.edu/~victor/papers/wtfa.pdf},
Author = {Nikolaj Hess-Nielsen and Mladen Victor Wickerhauser},
Title = {Wavelets and Time-Frequency Analysis},
Abstract = { We present a selective overview of time-frequency
analysis and some of its key problems. In particular we
motivate the introduction of wavelet and wavelet packet
analysis. Different types of decompositions of an
idealized time-frequency plane provide the basis for
understanding the performance of the numerical
algorithms and their corresponding interpretations
within the continuous models. As examples we show how
to control the frequency spreading of wavelet packets
at high frequencies using non-stationary filtering and
study some properties of periodic wavelet packets.
Furthermore we derive a formula to compute the time
localization of a wavelet packet from its indices which
is exact for linear phase filters, and show how this
estimate deteriorates with deviation from linear phase.},
Journal = {Proceedings of the IEEE},
Volume = {84},
Number = {4},
Pages = {523-540},
Note = {Special issue on wavelet applications},
Month = {April},
Year = {1996})
@InCollection(cw:wawd,
URL = {http://www.math.wustl.edu/~victor/papers/eeg.pdf},
Author = {Ronald Raphael Coifman and Mladen Victor Wickerhauser},
Title = {Wavelets, Adapted Waveforms, and De-Noising},
Abstract = {This is a short summary of a talk presented by
Wickerhauser at the Frontier Science in EEG Symposium,
``Continuous Waveform Analysis,'' held 9 October 1993
in New Orleans. We describe some new libraries of
waveforms well-adapted to various numerical analysis
and signal processing tasks. The main point is that by
expanding a signal in a library of waveforms which are
well-localized in both time and frequency, one can
achieve both understanding of structure and efficiency
in computation. We briefly cover the properties of the
new ``wavelet packet'' and ``localized trigonometric''
libraries. The main focus will be applications of such
libraries to the analysis of complicated transient
signals: a feature extraction and data compression
algorithm for speech signals which uses best-adapted
time and frequency decompositions, and an adapted
waveform analysis algorithm for removing fish noises
from hydrophone recordings. These signals share many
of the same properties as EEG traces, but with distinct
features that are easier to characterize and detect.},
BookTitle = {Continuous Wave-Form Analysis},
Editor = {Richard M. Dashieff and Diana J. Vincent},
Series = {Electroencephalography and Clinical Neurophysiology,
Supplement 45},
Pages = {57--78},
Publisher = {Elsevier},
Address = {New York},
Year = {1996})
@InProceedings(w:cwpicd,
URL = {http://www.math.wustl.edu/~victor/papers/cwpicd.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Custom Wavelet Packet Image Compression Design},
Abstract = {This tutorial paper presents a meta-algorithm for
designing a transform coding image compression
algorithm specific to a given application. The goal is
to select a decorrelating transform which performs best
on a given collection of data. It consists of
conducting experimental trials with adapted wavelet
transforms and the best basis algorithm, evaluating the
basis choices made for a training set of images, then
selecting a transform that, on average, delivers the
best compression for the data set. A crude version of
the method was used to design the WSQ fingerprint image
compression algorithm.},
BookTitle = {Proceedings of the 3rd International Workshop on Image and
Signal Processing, Manchester, UK, 4--7 November 1996},
Editor = {Todor Cooklev},
Publisher = {UMIST},
Organization = {UMIST},
Address = {Manchester, UK},
Pages = {7},
Year = {1996})
@InProceedings(w:bmw,
URL = {http://www.math.wustl.edu/~victor/papers/bmw.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Custom Wavelet Packet Image Compression for Multimedia},
Abstract = { We show how to design a transform coding image
compression algorithm specific to a given application.
A ``joint best basis'' decorrelating transform is
chosen which, on average, performs best on a given
collection of data. The transform always has low
complexity, and is adapted to the training set used to
choose it.},
BookTitle = {Tutorials of the Broadband and Multimedia Workshop,
Zagreb, Croatia},
Editor = {Mladen Kos},
Publisher = {University of Zagreb},
Organization = {FER},
Address = {Zagreb, Croatia},
Pages = {7},
Month = {11--12 November},
Year = {1996})
@Book(w:awatus,
Author = {Mladen Victor Wickerhauser},
Title = {Adaptive Wavelet-Analysis, theorie und software},
Translator = {Kurt Jetter},
Publisher = {Vieweg Verlag},
Address = {Braunschweig/Wiesbaden},
ISBN = {3-528-06688-1},
Pages = {xii + 440},
Note = {German translation of ``Adapted Wavelet Analysis from
Theory to Software''},
Month = {12 December},
Year = {1995})
@Article(w:mcc94,
URL = {http://www.math.wustl.edu/~victor/papers/mcc94.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Time Localization Techniques for Wavelet Transforms},
Abstract = { We consider the following pair of problems related to
orthonormal compactly supported wavelet expansions: (1)
Given a wavelet coefficient with its nominal scale and
position indices, find the precise location of the
transient signal feature which produced it; (2) Given
two collections of wavelet coefficients, determine
whether they arise from a periodic signal and its
translate, and if so find the translation which maps
one into the other. Both problems may be solved by
traditional means after inverting the wavelet
transform, but we propose two alternative algorithms
which rely solely on the wavelet coefficients
themselves.},
Note = {Proceedings of the Ninth {D}ubrovnik International
Course and {M}ath-{C}hem-{C}omp 1994},
Journal = {Croatica Chemica Acta},
Volume = {68},
Number = {1},
Pages = {1--27},
Month = {April},
Year = {1995})
@TechReport(tw:rbswetd,
URL = {http://www.math.wustl.edu/~victor/papers/tw.pdf},
Author = {Aurelija Trgo and Mladen Victor Wickerhauser},
Title = {A Relation between {S}hannon--{W}eaver Entropy and
``Theoretical Dimension'' for classes of Smooth functions},
Abstract = { Suppose that an infinite sequence is produced by
independent trials of a random variable with a fixed
distribution. The Shannon--Weaver entropy of the
sequence determines the minimum bit rate needed to
transmit the values of the sequence. We show that if
the source distribution is highly concentrated, as is
commonly observed in practice, then its entropy is
equal to the logarithm of the theoretical dimension of
the sequence. We conclude that the best-basis
algorithm, which minimizes this theoretical dimension
over a library of transformations, both chooses the
transformation that yields best compression and also
gives an estimate of the compression rate.},
Type = {Preprint},
Institution = {Washington University},
Address = {Saint Louis, Missouri},
Pages = {7},
Year = {1995})
@InProceedings(cmw:nha,
URL = {http://www.math.wustl.edu/~victor/papers/numharan.pdf},
Author = {Ronald Raphael Coifman and Yves Meyer and
Mladen Victor Wickerhauser},
Title = {Numerical Harmonic Analysis},
Abstract = { The purpose of this talk is to describe recent
developments involving the numerical implementation of
methods from classical harmonic analysis in signal
processing and computational P.D.E.},
Pages = {162--174},
BookTitle = {Essays on {F}ourier Analysis in Honor of
{E}lias {M.} {S}tein},
Editor = {Charles Fefferman and Robert Fefferman and Stephen Wainger},
ISBN = {0-691-08655-9},
Publisher = {Princeton University Press},
Address = {Princeton, New Jersey},
Note = {Proceedings of the Princeton Conference in
Harmonic Analysis, held 13--17 May 1991},
Year = {1995})
@Article(cw:emb,
URL = {http://www.math.wustl.edu/~victor/papers/emb.pdf},
Author = {Ronald Raphael Coifman and Mladen Victor Wickerhauser},
Title = {Adapted Waveform ``de-Noising'' for Medical Signals and
Images},
Abstract = {We describe some new libraries of waveforms
well-adapted to various numerical analysis and signal
processing tasks. The main point is that by expanding
a signal in a library of waveforms which are
well-localized in both time and frequency, one can
achieve both understanding of structure and efficiency
in computation. We briefly cover the properties of the
new ``wavelet packet'' and ``localized trigonometric''
libraries. The main focus will be applications of such
libraries to the analysis of complicated transient
signals: a feature extraction and data compression
algorithm for speech signals which uses best-adapted
time and frequency decompositions, and an adapted
waveform analysis algorithm for removing fish noises
from hydrophone recordings. These signals share many
of the same properties as EEG traces, but with distinct
features that are easier to characterize and detect.},
Journal = {IEEE Engineering in Medicine and Biology},
Volume = {14},
Number = {5},
Pages = {578--586},
Month = {September/October},
Year = {1995})
@Manual(w:awa3,
URL = {http://www.math.wustl.edu/~victor/papers/www.fmah.com},
Author = {Mladen Victor Wickerhauser},
Title = {AWA 3: Adapted Wavelet Analysis Library, version 3},
Note = {Software Documentation},
Organization = {Fast Mathematical Algorithms and Hardware Corporation},
Address = {Hamden, Connecticut},
Month = {June},
Year = {1995})
@InProceedings(w:spie2277,
URL = {http://www.math.wustl.edu/~victor/papers/timeloc.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Time Localization Techniques for Wavelet Transforms},
Abstract = { We consider the following pair of problems related
to orthonormal compactly supported wavelet expansions:
(1) Given a wavelet coefficient with its nominal scale
and position indices, find the precise location of the
transient signal feature which produced it; (2) Given
two collections of wavelet coefficients, determine
whether they arise from a periodic signal and its
translate, and if so find the translation which maps
one into the other. Both problems may be solved by
traditional means after inverting the wavelet
transform, but we propose two alternative algorithms
which rely solely on the wavelet coefficients
themselves.},
Pages = {18},
Editor = {Richard J. Mammone and J. David Murley Jr},
BookTitle = {Automatic Systems for the Identification
and Inspection of Humans},
Organization = {SPIE},
Publisher = {SPIE},
Series = {SPIE Proceedings},
Volume = {2277},
Address = {San Diego, California},
Month = {24--29 July},
Year = {1994})
@InCollection(w:cpcm,
URL = {http://www.math.wustl.edu/~victor/papers/taormina.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Comparison of Picture Compression Methods: Wavelet,
Wavelet Packet, and Local Cosine Transform Coding},
Abstract = { This summary describes several experiments in picture
compression using wavelets and the local cosine
transform of Coifman and Meyer. It describes an
adaptive wavelet transform coding method and a local
cosine transform algorithm based on the idea of a
``best basis,'' and provide Standard C algorithms for
computing some of the described transforms.},
Pages = {585--621},
Editor = {Charles K. Chui and Laura Montefusco and Luigia Puccio},
BookTitle = {Wavelets: Theory, Algorithms, and Applications},
Series = {Proceedings of the International Conference in
Taormina, Sicily, 14--20 October 1993},
Organization = {University of Messina},
Publisher = {Academic Press},
Address = {San Diego, California},
ISBN = {0-12-174575-9},
Year = {1994})
@InCollection(wfgwc:ecwpalc,
URL = {http://www.math.wustl.edu/~victor/papers/taormina2.pdf},
Author = {Mladen Victor Wickerhauser and Marie Farge and
Eric Goirand and Eva Wesfreid and Echeyde Cubillo},
Title = {Efficiency Comparison of Wavelet Packet and Adapted
Local Cosine Bases for Compression of a Two-dimensional
Turbulent Flow},
Abstract = { We compare the efficiency of two rank-reduction
methods for representing the essential features of a
two-dimensional turbulent vorticity field. The two
methods are both projections onto the largest
components, in one case onto the wavelet packet best
basis, in the other case onto the best basis of adapted
local cosines. We compare the two methods in three
ways: for efficiency of capturing enstrophy or
square-vorticity, for faithfulness to the power
spectrum, and for precision in resolving coherent
structures. These properties are needed for subsequent
segmentation into isolated coherent structures, or for
measurement of statistical quantities related to
coherent structures. We find that in all three
respects the wavelet packet representation is superior
to the local cosine representation.},
Pages = {509--531},
Editor = {Charles K. Chui and Laura Montefusco and Luigia Puccio},
BookTitle = {Wavelets: Theory, Algorithms, and Applications},
Series = {Proceedings of the International Conference in
Taormina, Sicily, 14--20 October 1993},
Organization = {University of Messina},
Publisher = {Academic Press},
Address = {San Diego, California},
ISBN = {0-12-174575-9},
Year = {1994})
@InProceedings(w:slob2,
URL = {http://www.math.wustl.edu/~victor/papers/slob.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Smooth Localized Orthonormal Bases},
Abstract = { We describe a decomposition of L^2(R) into an
orthogonal direct sum of copies of L^2(T). The
decomposition maps smooth functions to smooth periodic
functions. It generalizes certain earlier
constructions of smooth orthonormal windowed bases. In
particular, it shows the existence of smooth
orthonormal windowed exponential, wavelet, and wavelet
packet bases for L^2(R).},
Pages = {160--173},
Editor = {Alfred Z. Msezane and Katrina L. Barnum},
BookTitle = {Proceedings of the Sixth Annual Conference of the
National Alliance of Research Centers of Excellence},
Month = {17--19 March},
Year = {1994},
Address = {Clark Atlanta University, Atlanta, Georgia 30314},
Organization = {The Center for Theoretical Studies of Physical Systems})
@Article(w:waa,
URL = {http://www.math.wustl.edu/~victor/papers/meyer.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {{\bf Wavelets: Algorithms and Applications} by
{Y}ves {M}eyer},
Abstract = { Review of {\em Wavelets : Algorithms and Applications}, by
Yves Meyer, Society for Industrial and Applied Mathematics,
Philadelphia, 1993, softcover, pp. 133, \$19.50, ISBN
0-89871-309-9.},
Note = {Book review},
Pages = {526--528},
Journal = {SIAM Review},
Volume = {36},
Number = {526--528},
Month = {September},
Year = {1994})
@InProceedings(w:awfc,
URL = {http://www.math.wustl.edu/~victor/papers/awfc.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {An Adapted Waveform Functional Calculus},
Abstract = { We briefly survey how to use libraries of (orthonormal)
bases of well-behaved waveforms, including wavelets and
lapped orthogonal transforms, so as to obtain fast
numerical algorithms for the expansion of functions and
operators in these bases. The most important
applications are fast approximate matrix
multiplication, and application of matrices to vectors.},
Pages = {418--421},
BookTitle = {Proceedings of the Cornelius Lanczos Centenary,
Raleigh, North Carolina, 12--17 December 1993},
Editor = {Moody Chu and Robert Plemmons and David Brown and
Donald Ellison},
Publisher = {SIAM Press},
Organization = {SIAM},
ISBN = {0-89871-339-0},
Address = {Philadelphia},
Year = {1994})
@InProceedings(ww:tpom,
Author = {Eva Wesfreid and Mladen Victor Wickerhauser},
Title = {Traitement de la Parole par Ondelettes de {M}alvar},
BookTitle = {Reconnaisance Automatique de la Parole},
Editor = {J. P. Haton},
Series = {Actes du S{\'e}minaire},
Organization = {CRIN/INRIA--Nancy},
Month = {10--11 March},
Year = {1994})
@Article(cw:oe,
URL = {http://www.math.wustl.edu/~victor/papers/awaatr.pdf},
Author = {Ronald Raphael Coifman and Mladen Victor Wickerhauser},
Title = {Adapted Waveform Analysis as a Tool for Modeling,
Feature Extraction, and Denoising},
Abstract = { We describe the development of Adapted Waveform
Analysis (AWA) as a tool for fast processing of the
various identification tasks involved in medical
diagnostics and Automatic Target Recognition.},
Journal = {Optical Engineering},
Month = {July},
Volume = {33},
Number = {7},
Pages = {2170--2174},
Note = {Special issue on Adapted Wavelet Analysis},
Year = {1994})
@InCollection(cmqw:2,
URL = {http://www.math.wustl.edu/~victor/papers/cmqw.pdf},
Author = {Ronald Raphael Coifman and Yves Meyer and
Stephen R. Quake and Mladen Victor Wickerhauser},
Title = {Signal Processing and Compression with Wavelet Packets},
Abstract = { Some new algorithms for signal processing
and data compression arise from the discovery of certain
orthogonal functions which we shall call wavelet
packets. Wavelet packets generalize the compactly
supported wavelets of I. Daubechies and Y. Meyer. The
algorithms combine fast factored tranformations
with a tree-structure search for an optimal orthonormal
basis subset, some processing of the coefficients, and
then reconstruction of the transformed sequence.},
Editor = {James S. Byrnes and Jennifer L. Byrnes and
Kathryn A. Hargreaves and Karl Berry},
BookTitle = {Wavelets and Their Applications},
Publisher = {Kluwer Academic Publishers},
Address = {Dordrecht/Boston/London},
Series = {NATO ASI Series C: Mathematical and Physical Sciences},
Volume = {442},
Note = {Proceedings of the NATO Advanced Study Institute at
Il Ciocco, Barga, Italy in August, 1992},
Year = {1994},
ISBN = {0-7923-3078-1},
Pages = {363--379})
@InCollection(fgw:p2dwpt,
URL = {http://www.math.wustl.edu/~victor/papers/parallel.pdf},
Author = {Eric Goirand and Mladen Victor Wickerhauser and Marie Farge},
Title = {A Parallel Two Dimensional Wavelet Packet Transform and
Its Application to Matrix-Vector Multiplication},
Abstract = { This paper describes an implementation on an MIMD
computer of the 2-dimensional periodic wavelet packet
transform, the 2-dimensional ``best basis'' choice
algorithm, and the nonstandard matrix multiplication
algorithm. Our implementation also
works for the wavelet transform numerical algorithms of
Beylkin, Coifman and Rokhlin. It is one way to
obtain a fast functional calculus for certain classes
of linear operators: those operators, which sparsify in
the wavelet basis or the ``best basis'' of wavelet
packets, can be applied to vectors in a lower order of
complexity. The purpose of parallelizing the transform
is to distribute the cost of the initial sparsification
``investment'' over a large number of machines. This
one-time cost is O(N^2 log N) with a nonnegligible
constant; we envision applications in which N=10^6, for
example evolutions of 2-dimensional fluid velocity
fields on 1000x1000 point grids.
In a first part, we study a parallel algorithm (on a
MIMD machine) to compute the two-dimensional wavelet
packet transform. Then, we apply it to compute the
multiplication of a matrix by a vector in parallel.
We will compute matrix coefficients for operators with
respect to an orthonormal basis of separable wavelet
packets, using the 2-dimensional version of the fast
wavelet packet transform. The main idea is to
lift an NxN matrix, which maps R^N to R^N, into the
space of maps from R^{N log N} to R^{N log N}. Any of
a large number of NxN-coefficient subsets of this
lifting can be used to represent the operator, so we
may pick the subset in which the matrix is most sparse.
The choice is made with the ``best-basis'' algorithm
and is itself a fast algorithm. The number of
basissubsets of this type grows like 4^N for an NxN
matrix, but computing all of the coefficients with
respect to all of the basis functions requires just
O(N^2 log N) operations. The algorithm chooses
the ``best basis'' in which the operator appears most
sparse; it has complexity O(N^2).
We have reason to believe that the equations of motion
for important physical systems sparsify by nearly an
order of magnitude in this collection of bases. We
thus obtain lower complexity matrix application and
matrix multiplication algorithms from the new
representation. Of course the method is not perfectly
general: the speedup depends very much upon the
problem. However, others have shown that for broad
classes of operators we can expect an order of
magnitude asymptotic complexity reduction for matrix
application.},
BookTitle = {Wavelet Applications in Chemical Engineering},
Editor = {Rodolphe L. Motard and Babu Joseph},
Publisher = {Kluwer Academic Publishers},
Address = {Norwell, Massachusetts},
ISBN = {0-7923-9461-5},
Pages = {275--319},
Year = {1994})
@Article(w:mcc1,
URL = {http://www.math.wustl.edu/~victor/papers/rovinj.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Large-rank Approximate Principal Component Analysis with
Wavelets for Signal Feature Discrimination and the
Inversion of Complicated Maps},
Journal = {Journal of Chemical Information and Computer Science},
Volume = {34},
Abstract = { Principal orthogonal decomposition can be used to solve
two related problems: distinguishing elements from a
collection by making d measurements, and inverting a
complicated map from a p-parameter configuration space
to a d-dimensional measurement space. In the case
where d is more than 1000 or so, the classical O(d^3)
singular value decomposition algorithm becomes very
costly, but it can be replaced with an approximate
best-basis method that has complexity O(d^2\log d).
This can be used to compute an approximate Jacobian for
a complicated map from R^p to R^d in the case p < < d.},
Number = {5},
Month = {September/October},
Pages = {1036--1046},
Year = {1994})
@Book(w:awaftts,
Author = {Mladen Victor Wickerhauser},
Title = {Adapted Wavelet Analysis from Theory to Software},
URL = {http://www.math.wustl.edu/~victor/awaftts},
Abstract = { This text goes beyond the existing literature
to aid the engineer and applied mathematician in
writing computer programs to analyze real data. It
addresses the properties of wavelet and related
transforms, to establish criteria by which the proper
analysis tool may be chosen, and then details software
implementations to perform the needed computation. It
will also be useful to the pure mathematician who is
familiar with some parts of wavelet theory but has
questions about the applications. The worked exercises
make this a useful textbook for self-study, or for a
course in the theory and practice of wavelet analysis.
Beginning with an overview of the mathematical
prerequisites, successive chapters rigorously examine
the properties of the waveforms used in adapted wavelet
analysis: discrete ``fast'' Fourier transforms,
orthogonal and biorthogonal wavelets, wavelet packets,
and localized trigonometric or lapped orthogonal
functions. Other chapters discuss the ``best basis''
method, time-frequency analysis, and combinations of
these algorithms useful for signal analysis,
de-noising, and compression.
Each chapter discusses the technicalities of
implementation, giving examples in pseudocode backed up
with machine-readable Standard C source code available
on the optional diskette. Each chapter finishes with a
list of worked exercises in both the mathematics and
the programming of adapted wavelet algorithms.
Especially emphasized are the pitfalls and limitations
of the algorithms, with examples and suggestions given
to show how to avoid them.},
Publisher = {AK Peters, Ltd.},
Address = {Wellesley, Massachusetts},
ISBN = {1-56881-041-5},
Pages = {xii + 486},
Year = {1994})
@Article(w:oe,
URL = {http://www.math.wustl.edu/~victor/papers/oe.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Two Fast Approximate Wavelet Algorithms for Image
Processing, Classification, and Recognition},
Abstract = { We use large libraries of template waveforms with
remarkable orthogonality properties to recast the
relatively complex principal orthogonal decomposition
(POD) into an optimization problem with a fast solution
algorithm. Then it becomes practical to use POD to
solve two related problems: recognizing or classifying
images, and inverting a complicated map from a
low-dimensional configuration space to a
high-dimensional measurement space. In the case where
the number N of pixels or measurements is more than
1000 or so, the classical O(N^3) POD algorithm becomes
very costly, but it can be replaced with an approximate
best-basis method that has complexity O(N^2 log N). A
variation of POD can also be used to compute an
approximate Jacobian for the complicated map.},
Journal = {Optical Engineering},
Month = {July},
Volume = {33},
Number = {7},
Pages = {2225--2235},
Note = {Special issue on Adapted Wavelet Analysis},
Year = {1994})
@InProceedings(w:spie2242,
URL = {http://www.math.wustl.edu/~victor/papers/spie2242.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Wavelet Approximations to {J}acobians and the Inversion
of Complicated Maps},
Editor = {Harold H. Szu},
BookTitle = {Wavelet Applications},
Organization = {SPIE},
Publisher = {SPIE},
Series = {SPIE Proceedings},
Volume = {2242},
ISBN = {0-8194-1546-4},
Address = {Orlando, Florida},
Month = {5--8 April},
Pages = {100--118},
Year = {1994})
@Article(ww:alttsp,
URL = {http://www.math.wustl.edu/~victor/papers/alttasp.pdf},
Author = {Eva Wesfreid and Mladen Victor Wickerhauser},
Title = {Adapted Local Trigonometric Transform and Speech Processing},
Abstract = { We use an algorithm based on the adapted-window
Malvar transform to decompose digitized speech signals
into a local time-frequency representation. We present
some applications and experimental results for signal
compression and automatic voiced-unvoiced
segmentation. This decomposition provides a method of
parameter simlification which appears to be useful for
detecting fundamental frequencies and characterizing
formants. Note: Additional figures are in sono1.eps
and sono2.eps in the file archive.},
Journal = {IEEE Transactions on Signal Processing},
Number = {12},
Volume = {41},
Pages = {3596--3600},
Month = {December},
Year = {1993})
@Misc(ow:wsq,
Title = {{WSQ} -- the {FBI}/{Y}ale/{L}os {A}lamos [{W}]avelet-packet
[{S}]calar [{Q}]uantization fingerprint compression
algorithm, for {W}indows 3.1 or higher},
Author = {He Ouyang and Mladen Victor Wickerhauser},
HowPublished = {Noncompliant with latest standard---no longer available.
See the certified codes
elsewhere at this site.},
Year = {1993},
Month = {8 September})
@InProceedings(afww,
Author = {Christophe D'Alessandro and Xiang Fang and Eva Wesfreid
and Mladen Victor Wickerhauser},
Title = {Speech Signal Segmentation via {M}alvar Wavelets},
Pages = {305--308},
BookTitle = {Progress in Wavelet Analysis and Applications},
Editor = {Yves Meyer and Sylvie Roques},
Series = {Proceedings of the International Conference ``Wavelets
and Applications,'' Toulouse, France, 8--13 June 1992},
Publisher = {Editions Frontieres},
Organization = {Observatoire Midi-Pyr{\'e}n{\'e}es de l'Universit{\'e}
Paul Sabatier},
Address = {Gif-sur-Yvette, France},
ISBN = {2-86332-130-7},
Year = {1993})
@InCollection(cmqw:3,
URL = {http://www.math.wustl.edu/~victor/papers/cmqw.pdf},
Author = {Ronald Raphael Coifman and Yves Meyer and
Stephen R. Quake and Mladen Victor Wickerhauser},
Title = {Signal Processing and Compression with Wavelet Packets},
Abstract = { We describe some new algorithms for signal processing
and data compression based on a collection of
orthogonal functions which we shall call wavelet
packets. Wavelet packets generalize the compactly
supported wavelets of I. Daubechies and Y. Meyer. The
algorithms we describe combine the projection of a
sequence onto wavelet packet components, the selection
of an optimal orthonormal basis subset, some linear or
quasilinear processing of the coefficients, and then
reconstruction of the transformed sequence.},
Pages = {77--93},
BookTitle = {Progress in Wavelet Analysis and Applications},
Editor = {Yves Meyer and Sylvie Roques},
Series = {Proceedings of the International Conference ``Wavelets and
Applications,'' Toulouse, France, 8--13 June 1992},
Publisher = {Editions Frontieres},
Organization = {Observatoire Midi-Pyr{\'e}n{\'e}es de l'Universit{\'e}
Paul Sabatier},
Address = {Gif-sur-Yvette, France},
ISBN = {2-86332-130-7},
Year = {1993})
@InCollection(cw:d-psam,
URL = {http://www.math.wustl.edu/~victor/papers/wawa.pdf},
Author = {Ronald Raphael Coifman and Mladen Victor Wickerhauser},
Title = {Wavelets and Adapted Waveform Analysis: A Toolkit for
Signal Processing and Numerical Analysis.},
Abstract = {Our goal is to describe tools for adapting methods
of analysis to various tasks occurring in harmonic and
numerical analysis and signal processing. The main
point of this presentation is that by choosing an
orthonormal basis, in which space and frequency are
suitably localized, one can achieve both understanding
of structure and efficiency in computation. We
describe a fingerprint image segmentation algorithm, an
alternative factorization for the FFT, and a
wavelet-based denoising algorithm.},
Note = {Minicourse lecture notes},
Pages = {119--153},
BookTitle = {Different Perspectives on Wavelets},
Series = {Proceedings of Symposia in Applied Mathematics},
Number = {47},
Editor = {Ingrid Daubechies},
ISBN = {0-8218-5503-4},
Publisher = {American Mathematical Society},
Address = {San Antonio, Texas},
Month = {11-12 January},
Year = {1993})
@InCollection(w:bawpb,
URL = {http://www.math.wustl.edu/~victor/papers/bestbase.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Best-adapted Wavelet Packet Bases},
Abstract = { This paper is a review of the construction of
orthogonal wavelet packets, using the quadrature mirror
filter algorithm slightly generalized to the case of
p>2 or p=2 wavelets and scaling functions. It is part
of the AMS short course on ``Wavelets and
Applications'' held in San Antonio, 11-12 January 1993.},
Note = {Minicourse lecture notes},
Pages = {155--171},
BookTitle = {Different Perspectives on Wavelets},
Series = {Proceedings of Symposia in Applied Mathematics},
Number = {47},
Editor = {Ingrid Daubechies},
ISBN = {0-8218-5503-4},
Publisher = {American Mathematical Society},
Address = {San Antonio, Texas},
Month = {11-12 January},
Year = {1993})
@Article(w:slob,
URL = {http://www.math.wustl.edu/~victor/papers/crasslob.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Smooth Localized Orthonormal Bases},
Abstract = { We describe an orthogonal decomposition of
square-integrable functions on the line, which maps
smooth functions to smooth periodic functions. It
generalizes previous constructions by Malvar, Coifman
and Meyer. The adjoint of the decomposition can be
used to construct smooth orthonormal windowed
exponential, wavelet and wavelet packet bases.},
Journal = {Comptes Rendus de l'Acad{\'e}mie des Sciences de Paris},
Series = {I},
Volume = {316},
Pages = {423--427},
Year = {1993})
@InProceedings(w:cwatfa,
URL = {http://www.math.wustl.edu/~victor/papers/cwatfa.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Computation with Adapted Time-Frequency Atoms},
Abstract = { Operators can be represented by their coefficients
with respect to an orthonormal basis of functions well
localized in both time and frequency. Such building
blocks will be called time-frequency atoms. Bases of
functions of this type abound and one may be chosen to
minimize the number of nonnegligible coefficients for a
given operator. In this note we will use particular
libraries of such atoms to represent operators
efficiently and thereby obtain a ``fast'' functional
calculus. We obtain lower complexity matrix
application and matrix multiplication algorithms from
the new representation. The main idea is to lift an
NxN matrix into the space of maps on (N log N)x(N log
N). Any of a large number of NxN-coefficient subsets
of this lifting can be used to represent the operator,
so we may pick the subset in which the matrix is most
sparse. The choice is made with the ``best-basis''
algorithm and is itself a fast algorithm. The
best-basis implementations described here can use
either wavelet packets or adapted local trigonometric
libraries. They provide a generalization of the
``nonstandard form'' introduced by Beylkin, Coifman,
and Rokhlin.},
Pages = {175--184},
BookTitle = {Progress in Wavelet Analysis and Applications},
Editor = {Yves Meyer and Sylvie Roques},
Series = {Proceedings of the International Conference ``Wavelets
and Applications,'' Toulouse, France, 8--13 June 1992},
Publisher = {Editions Frontieres},
Organization = {Observatoire Midi-Pyr{\'e}n{\'e}es de l'Universit{\'e}
Paul Sabatier},
Address = {Gif-sur-Yvette, France},
ISBN = {2-86332-130-7},
Year = {1993})
@Misc(wplab,
URL = {http://www.math.wustl.edu/~victor/software/WPLab/WPLab3.app.tar.gz},
Title = {{WPL}ab version 3 (for {NeXT} computers)},
Author = {David Rochberg and Mladen Victor Wickerhauser},
HowPublished = {Available by anonymous file transfer},
Year = {1993})
@InProceedings(cmw:iciam91,
URL = {http://www.math.wustl.edu/~victor/papers/adaptwave1.pdf},
Title = {Adapted Waveform Analysis, Wavelet-Packets and Applications},
Author = {Ronald Raphael Coifman and Yves Meyer and
Mladen Victor Wickerhauser},
Abstract = { Adapted wave form analysis, refers to a collection of
FFT like adapted transform algorithms. Given a
function or an operator these methods provide a special
orthonormal basis relative to which the function is
well represented, and the operator is described by a
sparse matrix. The selected basis functions are chosen
inside predefined libraries of oscillatory localized
functions (waveforms). These algorithms are of
complexity N log N, opening the door for a large range
of applications in signal and image processing, as well
as in numerical analysis. Our goal is to describe and
relate traditional windowed Fourier Transform methods
to wavelet, wavelet-packet base algorithms by making
explicit their dual nature and relative role in
analysis and computation.},
Pages = {41--50},
BookTitle = {ICIAM 91: Proceedings of the 2nd International Conference
on Industrial and Applied Mathematics, 8--12 July, 1991},
Editor = {Robert E. O'Malley Jr.},
Publisher = {SIAM Press},
ISBN = {0-89871-302-1},
Organization = {SIAM},
Address = {Philadelphia},
Year = {1992})
@InCollection(aww,
URL = {http://www.math.wustl.edu/~victor/papers/aww.pdf},
Author = {Pascal Auscher and Guido Leopold Weiss and
Mladen Victor Wickerhauser},
Title = {Local Sine and Cosine Bases of {C}oifman and {M}eyer
and the Construction of Smooth Wavelets},
Abstract = { We give a detailed account of the local cosine and
sine bases of Coifman and Meyer. We describe some of
their applications; in particular, based on an approach
by Coifman and Meyer, we show how these local bases can
be used to obtain arbitrarily smooth wavelets. The
understanding of this material requires only a minimal
knowledge of the Fourier transform and classical
analysis. It is our intention to make this
presentation accessible to all who are interested in
Wavelets and their applications.},
Pages = {237--256},
Editor = {Charles K. Chui},
BookTitle = {Wavelets--A Tutorial in Theory and Applications},
Publisher = {Academic Press},
Address = {Boston},
ISBN = {0-12-174590-2},
Year = {1992})
@InCollection(cmw:spwp,
URL = {http://www.math.wustl.edu/~victor/papers/sizeprop.pdf},
Author = {Ronald Raphael Coifman and Yves Meyer and
Mladen Victor Wickerhauser},
Title = {Size Properties of Wavelet Packets},
Abstract = { We investigate the frequency localization of wavelet
packets and prove that they do not enjoy sharp
frequency localization. By S. Bernstein's
inequalities, a sharp frequency localization would
imply a uniform bound on the supremum norm of the basic
wavelet-packets w_n(x). But theorem 3 shows that the
average growth of ess sup |w_n| is n^gamma for some
positive gamma. The fact that gamma is rather small
plays a key role in the construction of a large library
of wavelet packet orthonormal bases.},
Pages = {453--470},
BookTitle = {Wavelets and Their Applications},
Editor = {Mary Beth Ruskai and Gregory Beylkin and Ronald Raphael
Coifman and Ingrid Daubechies and St{\'e}phane Mallat
and Yves Meyer and Louise Raphael},
Publisher = {Jones and Bartlett},
Address = {Boston},
ISBN = {0-86720-225-4},
Year = {1992})
@InCollection(cmw:wasp,
URL = {http://www.math.wustl.edu/~victor/papers/wasp.pdf},
Author = {Ronald Raphael Coifman and Yves Meyer and
Mladen Victor Wickerhauser},
Abstract = { Wavelet Analysis consists of a versatile collection of
tools for the analysis and manipulation of signals such
as sound and images,as well as more general digital
data sets. The user is provided with a collection of
standard libraries of waveforms, which can be chosen to
fit specific classes of signals. These libraries come
equipped with fast numerical algorithms enabling
realtime implementation of a variety of signal
processing tasks, such as data compression, extraction
of parameters for recognition and diagnostics,
transformation and manipulation of data. The process
of analysis of data is usually started by comparing
acquired segments of data with stored waveforms.},
Title = {Wavelet Analysis and Signal Processing},
Pages = {153--178},
BookTitle = {Wavelets and Their Applications},
Editor = {Mary Beth Ruskai and Gregory Beylkin and Ronald Raphael
Coifman and Ingrid Daubechies and St{\'e}phane Mallat
and Yves Meyer and Louise Raphael},
Publisher = {Jones and Bartlett},
Address = {Boston},
ISBN = {0-86720-225-4},
Year = {1992})
@InCollection(hpw,
URL = {http://www.math.wustl.edu/~victor/papers/burgers.pdf},
Author = {Fr{{\'e}}d{{\'e}}ric Heurtaux and Fabrice Planchon and
Mladen Victor Wickerhauser},
Title = {Scale Decomposition in {B}urgers' equation},
Abstract = { The wavelet representation of a time-dependent
signal can be used to study the propagation of energy
between the different scales in the signal. Burgers'
evolution operator (in 1 and 2 dimensions) can itself
be described >from this scaling point of view. Using
wavelet-based algorithms we can depict the transfer of
energy between scales. We can write the instantaneous
evolution operator in the wavelet basis; then large
off-diagonal terms will correspond to energy transfers
between different scales. We can project the solution
onto each fixed-scale wavelet subspace and compute the
energy; then the rate of change of this energy by scale
can detect and quantify any cascades that may be
present. These methods improve the classical
Fourier-transform-based scale decomposition which uses
the notion that wavenumber=scale. The wavelet basis
functions underlying our scale decompositions have
finite, well-defined position uncertainty (i.e., scale)
whereas Fourier basis functions have formally unbounded
position uncertainty.},
Pages = {505--523},
Editor = {John J. Benedetto and Michael Frazier},
BookTitle = {Wavelets: Mathematics and Applications},
Series = {Studies in Advanced Mathematics},
Publisher = {CRC Press},
Address = {Boca Raton, Florida},
ISBN = {0-8493-8271-8},
Year = {1992})
@InCollection(cw:wawa,
URL = {http://www.math.wustl.edu/~victor/papers/wawa.pdf},
Author = {Ronald Raphael Coifman and Mladen Victor Wickerhauser},
Title = {Wavelets and Adapted Waveform Analysis},
Abstract = {Our goal is to describe tools for adapting methods of
analysis to various tasks occuring in harmonic and
numerical analysis and signal processing. The main
point of this presentation is that by choosing an
orthonormal basis, in which space and frequency are
suitably localized, one can achieve both understanding
of structure and efficiency in computation. We
describe a fingerprint image segmentation algorithm, an
alternative factorization for the FFT, and a
wavelet-based denoising algorithm.},
Pages = {399--423},
Editor = {John J. Benedetto and Michael Frazier},
BookTitle = {Wavelets: Mathematics and Applications},
Series = {Studies in Advanced Mathematics},
Publisher = {CRC Press},
Address = {Boca Raton, Florida},
ISBN = {0-8493-8271-8},
Year = {1992})
@Article(cw:ebafbbs,
URL = {http://www.math.wustl.edu/~victor/papers/ebafbbs.pdf},
Author = {Ronald Raphael Coifman and Mladen Victor Wickerhauser},
Title = {Entropy Based Algorithms for Best Basis Selection},
Abstract = { We would like to describe a method permitting efficient
compression of a variety of signals such as sound and
images. While similar in goals to vector quantization,
the new method uses a codebook or library of predefined
modulated waveforms with some remarkable orthogonality
properties. We can apply the method to two
particularly useful libraries of recent vintage,
orthogonal wavelet-packets and localized trigonometric
functions, for which the time-frequency localization
properties of the waveforms are reasonably well
controlled. The idea is to build out of the library
functions an orthonormal basis relative to which the
given signal or collection of signals has the lowest
information cost. We may define several useful cost
functionals; one of the most attractive is Shannon
entropy, which has a geometric interpretation in this
context.
Practicality is built into the foundation of this
orthogonal best-basis methods. All bases from each
library of waveforms described below come equipped with
fast O(N log N) transformation algorithms, and each
library has a natural dyadic tree structure which
provides O(N log N) search algorithms for obtaining the
best basis. The libraries are rapidly constructible,
and never have to be stored either for analysis or
synthesis. It is never necessary to construct a
waveform from a library in order to compute its
correlation with the signal. The waveforms are indexed
by three parameters with natural interpretations
(position, frequency, and scale), and we have
experimented with feature-extraction methods that use
best-basis compression for front-end complexity
reduction.
The method relies heavily on the remarkable
orthogonality properties of the new libraries. It is
obviously a nonlinear transformation to represent a
signal in its own best basis, but since the
transformation is orthogonal once the basis is chosen,
compression via the best-basis method is not
drastically affected by noise: the noise energy in the
transform values cannot exceed the noise energy in the
original signal. Furthermore, we can use information
cost functionals defined for signals with normalized
energy, since all expansions in a given library will
conserve energy. Since two expansions will have the
same energy globally, it is not necessary to normalize
expansions to compare their costs. This feature
greatly enlarges the class of functionals usable by the
method, speeds the best-basis search, and provides a
geometric interpretation in certain cases.},
Journal = {IEEE Transactions on Information Theory},
Volume = {32},
Pages = {712--718},
Month = {March},
Year = {1992})
@Article(fgmpw,
URL = {http://www.math.wustl.edu/~victor/papers/fgmpw.tar.gz},
Author = {Marie Farge and Eric Goirand and Yves Meyer and
Fr{\'e}d{\'e}ric Pascal and Mladen Victor Wickerhauser},
Title = {Improved Predictability of Two-Dimensional Turbulent
Flows Using Wavelet Packet Compression},
Abstract = { We propose to use new orthonormal wavelet packet
bases, more efficient than the Fourier basis, to
compress two-dimensional turbulent flows. We define
the ``best basis'' of wavelet packets as the one which,
for a given enstrophy density, condenses the $L^2$ norm
into a minimum number of non-negligible wavelet packet
coefficients. Coefficents below a threshold are
discarded, reducing the number of degrees of freedom.
We then compare the predictability of the original flow
evolution with several such reductions, varying the
number of retained coefficients, either from a Fourier
basis, or from the best-basis of wavelet packets. We
show that for a compression ratio of 1/2, we still have
a deterministic predictability using the wavelet packet
best-basis, while it is lost when using the Fourier
basis. Likewise, for compression ratios of 1/20 and
1/200 we still have statistical predictability using
the wavelet packet best-basis, while it is lost when
using the Fourier basis. In fact, the significant
wavelet packet coefficients in the best-basis appear to
correspond to coherent structures. The weak
coefficients correspond to vorticity filaments, which
are only passively advected by the coherent structures.
In conclusion, the wavelet packet best-basis seems to
distinguish the low-dimensional dynamically active part
of the flow from the high-dimensional passive
components. It gives us some hope of drastically
reducing the number of degrees of freedom, which varies
as Reynolds, in the computation of two-dimensional
turbulent flows.},
Journal = {Fluid Dynamics Research},
Volume = {10},
Pages = {229--250},
Year = {1992})
@InCollection(w:acoustic,
URL = {http://www.math.wustl.edu/~victor/papers/acoustic.ps.gz},
Author = {Mladen Victor Wickerhauser},
Title = {Acoustic signal compression with wavelet packets},
Abstract = {The wavelet transform generalizes to produce a library
of orthonormal bases of modulated wavelet packets.
Each basis comes with a fast transform; these bases are
similar to adapted windowed Fourier transforms. There
is a notion of the ``best basis'' for a signal, given a
cost function. This paper discusses some early results
in acoustic signal compression using a simple counting
cost function.},
Pages = {679--700},
Editor = {Charles K. Chui},
BookTitle = {Wavelets--A Tutorial in Theory and Applications},
Publisher = {Academic Press},
Address = {Boston},
ISBN = {0-12-174590-2},
Year = {1992})
@Article(w:dsp,
URL = {http://www.math.wustl.edu/~victor/papers/dsp.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {High-Resolution Still Picture Compression},
Abstract = { We consider the problem of storing, transmitting, and
manipulating digital electronic images. Because of the
file sizes involved, transmitting images will always
consume large amounts of bandwidth, and storing images
will always require hefty resources. Because of the
large number N of pixels in a high resolution image,
manipulation of digital images is infeasible without
low-complexity algorithms, i.e., O(N) or O(N log(N)).
Our goal will be to describe some new methods which are
firmly grounded in harmonic analysis and the
mathematical theory of function spaces, which promise
to combine effective image compression with
low-complexity image processing. We shall take a broad
perspective, but we shall also compare specific new
algorithms to the state of the art.},
Month = {October},
Volume = {2},
Number = {4},
Pages = {204--226},
Journal = {Digital Signal Processing: a Review Journal},
Year = {1992})
@Manual(w:awa2,
URL = {http://www.fmah.com},
Author = {Mladen Victor Wickerhauser},
Title = {Adapted Waveform Analysis Library, v2.0},
Note = {Software Documentation},
Organization = {Fast Mathematical Algorithms and Hardware Corporation},
Address = {Hamden, Connecticut},
Month = {June},
Year = {1992})
@InCollection(zsw,
Author = {Lareef Zubair and Kannan R. Sreenivasan and
Mladen Victor Wickerhauser},
Title = {Characterization and Compression of Turbulent Signals
and Images using Wavelet Packets},
BookTitle = {Studies in Turbulence},
Pages = {489--513},
Editor = {T. Gadsky and S. Sirkar and C. Speziale},
Publisher = {Springer Verlag},
Address = {New York},
Year = {1991})
@InCollection(w:inria,
URL = {http://www.math.wustl.edu/~victor/papers/lwpa.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {{INRIA} Lectures on Wavelet Packet Algorithms},
Abstract = { We begin by defining continuous wavelet packets on R.
These are square-integrable functions with prescribed
smoothness and other properties, which we shall develop
to establish the main notions. Our construction will
be directed toward numerical applications, so we will
restrict ourselves to the quadrature mirror filter
algorithm.
Next we will define several discrete algorithms and
explore their advantages and disadvantages. We will
show the correspondence between wavelet packets and
coefficients computed from sampled signals, and relate
the convergence of this approximation to the smoothness
of the signal. We will define information cost
functions and the ``best-basis'' method. We will count
operations and consider practical matters like the
memory requirements of the algorithms, periodizing, the
spreading of the support of aperiodic wavelet packets,
and the combinatorics of constructing wavelet packet
bases of increasing generality.
In parallel, we will develop smooth orthogonal local
trigonometric transforms. These are properly
considered transposes of wavelet packet methods, or
alternatively conjugates of wavelet packet methods by
the Fourier transform. We will describe both
continuous and discrete local cosine transforms, and an
adaptive local cosine transform useful for signal
segmentation.
We will examine several compression methods, both
linear and nonlinear. Linear methods include uniform
and nonuniform quantization. Nonlinear methods include
discarding small coefficients, coalescing to the center
of energy within bands, and Karhunen--Loeve methods.
We will examine the peculiarities of each method, and
discuss the errors in the lossy versions of these
algorithms. This will illustrate the relative
advantages of wavelet packets, windowed Fourier
transforms, and wavelet bases.
We will then generalize to multidimensions by
separation of variables. We will explore the
combinatorics of higher dimensional wavelets. We will
identify matrices with two-dimensional signals or
``pictures,'' and we will show how each picture
compression algorithm yields a nonstandard matrix
multiplication algorithm.
As a demonstration of the analytic power of best-basis
methods, we will perform an automatic analysis of a few
canonical signals in the phase plane. The signals will
be decomposed into as precise a set of modulated lumps
as the Heisenberg uncertainty principle allows, and the
product of the analysis will be displayed in an
intuitively satisfying manner.
Finally, we will produce several fast numerical
algorithms driven by the best-basis method. Among
these will be fast approximate Karhunen--Loeve factor
analysis, signal segmentation in time and frequency,
feature-preserving encryption, and matrix
multiplication.},
Editor = {Pierre-Louis Lions},
BookTitle = {Probl{\'e}mes Non-Lin{\'e}aires Appliqu{\'e}s, Ondelettes
et Paquets D'Ondes},
Note = {Minicourse lecture notes},
Publisher = {INRIA},
Address = {Roquencourt, France},
Month = {17--21 June},
Pages = {31-99},
Year = {1991})
@InProceedings(w:fakle,
URL = {http://www.math.wustl.edu/~victor/papers/fakle.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Fast Approximate Factor Analysis},
Abstract = { The principal orthogonal factor analysis or
Karhunen-Loeve algorithm may be sped up by a
low-complexity preprocessing step. A fast transform is
selected from a large library of wavelet-like
orthonormal bases, so as to maximize transform coding
gain for an ensemble of vectors. Only the top few
coefficients in the new basis, in order of variance
across the ensemble, are then decorrelated by
diagonalizing the autocovariance matrix. The method
has computational complexity O(d*d*log d+ d'*d'*d') and
O(d*log d + d'*d') respectively for training and
classifying a $d$-dimensional system, where d'< < d.
One application is described, the reduction of an
ensemble of 16,384-pixel face images to a 10-parameter
space using a desktop computer, retaining 90 percent of
the variance of the ensemble.},
Pages = {23--32},
Editor = {Martine J. Silbermann and Hemant D. Tagare},
BookTitle = {Curves and Surfaces in Computer Vision and Graphics {II}},
Organization = {SPIE},
Series = {SPIE Proceedings},
Volume = {1610},
Pages = {iii + 395},
ISBN = {0-8194-0747-X},
Address = {Boston},
Month = {October},
Year = {1991})
@Book(cw:mmnotes,
URL = {http://www.fmah.com},
Author = {Ronald Raphael Coifman and Mladen Victor Wickerhauser},
Title = {{M}artin--{M}arietta Wavelet Lectures},
Publisher = {Martin--Marietta Corporation},
Month = {21--25 October},
Year = {1991})
@Misc(fgppw:video,
Author = {Marie Farge and Eric Goirand and Thierry Philipovitch and
Fre{\'e}d{\'e}ric Pascal and Mladen Victor Wickerhauser},
Title = {Wavelet Packets Compression of a 2D Turbulent Flow},
HowPublished = {Video recording of a computer simulation performed at
LMD-CNRS, Paris},
Year = {1991})
@Misc(wplab1,
URL = {http://www.math.wustl.edu/~victor/software/WPLab/WPLab1.5.gz},
Title = {{WPL}ab version 1.5 (for {NeXT} computers)},
Author = {Mladen Victor Wickerhauser},
HowPublished = {Available by anonymous file transfer},
Month = {6 April},
Year = {1991})
@Article(cw:bo,
URL = {http://www.math.wustl.edu/~victor/papers/benj-ono.pdf},
Author = {Ronald Raphael Coifman and Mladen Victor Wickerhauser},
Title = {The Scattering Transform for the
{B}enjamin--{O}no Equation},
Abstract = { We use constructive methods to investigate the spectral
theory of the Benjamin--Ono equation. Since the
linearization series used previously is singular, we
replace it with an improved series obtained by
finite-rank renormalization. This introduces additional
scattering data, which are shown to be dependent upon a
single function, though not the usual one. We then
prove the continuity of the direct and inverse
scattering transforms defined by the improved series
for small complex potentials. For all such potentials,
the eigenvalues of the spectral problem cannot
accumulate at 0. Rapidly decaying potentials have
regular scattering data, prohibiting rapidly decaying
solitons. In the selfadjoint case (real potentials),
we obtain explicit cancellation of certain
singularities. This leads to an alternate existence
proof for the Cauchy problem for the equation. It also
proves existence and gives estimates for some
previously formal invariant quantities associated to
the Benjamin--Ono hierarchy.},
Journal = {Inverse Problems},
Volume = {6},
Pages = {825--861},
Year = {1990})
@TechReport(w:pic,
URL = {http://www.math.wustl.edu/~victor/papers/pic.tar.gz},
Author = {Mladen Victor Wickerhauser},
Title = {Picture Compression by Best-Basis Sub-Band Coding},
Abstract = { We introduce a generalization of sub-band coding which
may be used to compress digitized pictures or sequences
of pictures. The method selects a most efficient
orthogonal representation of the picture from among a
large number of possibilities. The efficiency
functional need only be additive across direct sum
decompositions. We present some results of the method
using Shannon entropy as the efficiency functional, and
mean-square deviation as the error criterion.},
Type = {Preprint},
Institution = {Yale University},
Year = {1990})
@TechReport(w:nsmult,
URL = {http://www.math.wustl.edu/~victor/papers/nsmatrix.pdf},
Author = {Mladen Victor Wickerhauser},
Title = {Nonstandard Matrix Multiplication},
Abstract = { We describe an algorithm for multiplying matrices in
the compressed coordinates obtained by wavelet packet
transforms. This generalizes the ``nonstandard matrix
multiplication'' of Belykin, Coifman, and Rokhlin.},
Type = {Preprint},
Institution = {Yale University},
Month = {15 May},
Year = {1990})
@TechReport(gw:elemwave,
URL = {http://www.math.wustl.edu/~victor/papers/elemwave.pdf},
Author = {Elliot Gootman and Mladen Victor Wickerhauser},
Title = {Elementary Wavelets},
Abstract = { Necessary and sufficient conditions are given for some
functions in L^2(R) with compactly supported Fourier
transforms to generate orthonormal bases under the
action of integer translations and power-of-two
dilations.},
Type = {Preprint, 02021-88},
Institution = {Mathematical Sciences Research Institute},
Address = {Berkeley, California},
Year = {1988})
@Article(w:kp,
Author = {Mladen Victor Wickerhauser},
Title = {Hamilton's Form for the
{K}adomtsev--{P}etviashvili Equation},
Abstract = { The scattering transform for the
Kadomtsev--Petviashvili equation (KP-II) is a local
symplectomorphism. Pulling back the Hamiltonians for
the linear evolutions of scattering data gives
Hamiltonians for the KP-II hierarchy: they are values
of the associated scattering data at distinguished
points. This method yields simple proofs that KP-II
has infinitely many commuting flows and simplifies
their calculation. It also provides a Plancherel-type
theorem.},
Journal = {Journal of Mathematical Physics},
Volume = {29},
Pages = {2300--2302},
Year = {1988})
@Article(w:cmp,
Author = {Mladen Victor Wickerhauser},
Title = {Inverse Scattering for the Heat Operator
and Evolutions in 2+1 Variables},
Abstract = { The asymptotic behavior of functions in the kernel of
the perturbed heat operator D^2_x - D_y - u(x,y)
suffice to determine u. An explicit formula is derived
using the d-bar method of inverse scattering, complete
with estimates for small and moderately regular
potentials u. If u evolves so as to satisfy the
Kadomtsev-Petviashvili (KP-II) equation, the asymptotic
data evolve linearly and boundedly. Thus the KP-II
equation has solutions bounded for all time. A method
for calculating nonlinear evolutions related to KP-II
is presented. The related evolutions include the
so-called ``KP-II Hierarchy'' and many others.},
Journal = {Communications in Mathematical Physics},
Volume = {108},
Pages = {67--89},
Year = {1987})
@PhDThesis(w:phd,
Author = {Mladen Victor Wickerhauser},
Title = {Nonlinear Evolutions of the Heat Operator},
Abstract = { Scattering data, which is the asymptotic behavior of
solutions q = q(x,y) to the perturbed heat equation
D^_x q - Dy q = uq, is computed, and shown to determine
the perturbing function u=u(x,y). An explicit formula
is derived using the d-bar method of inverse
scattering, complete with estimates for small and
moderately regular u. A method of calculating
nonlinear evolutions of u which correspond to linear
evolutions of the scattering data is presented, and
shown to generate the Kadomtsev--Petviashvili II
hierarchy of evolutions.},
School = {Yale University},
Address = {New Haven, Connecticut},
Month = {May},
Year = {1985})
@TechReport(w:cyber,
Author = {Mladen Victor Wickerhauser},
Title = {{C}yber 2xx Performance on an Implicit Factored
{N}avier--{S}tokes Algorithm},
Abstract = { By modeling the execution of a particular numerical
algorithm, we determine that a proposed supercomputer
architecture is capable of nearly one billion
floating-point operations per second.},
Institution = {NASA Ames Research Center},
Type = {Preprint},
Year = {1981})
Questions? Return to
M. Victor Wickerhauser's home page for contact information.
Last modified on October 23, 2002.